This Week in WordPress #367 - CourtneyR.dev's 2nd Brain

This Week in WordPress #367 ![rw-book-cover](https://img.youtube.com/vi/VMzU4GhuuCI/maxresdefault.jpg) 🔗 **Source:** [This Week in WordPress #367](https://www.youtube.com/watch?v=VMzU4GhuuCI) > [!note]+ Document Notes > tags: [wordpress, core-development, security, performance, ai, open-source] > author: Nathan Wrigley > source: WP Builds > reference: https://www.youtube.com/watch?v=VMzU4GhuuCI > date created: 2025-02-24 14:00 > > # 🏷️ Title: This Week in WordPress #367 >  ## 🧰 Action Inbox - [H1] Test WordPress 7.0 Beta 1 on non-production site — collaborative editing features #wordpress - [H2] Consider open source AI models for text-to-speech accessibility — cost savings #accessibility - [H3] Implement password manager if not using one — security necessity #security ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) ### 🔗 References - [WordPress 7.0 Beta 1](https://wordpress.org/news/) - [MainWP Year in Review](https://mainwp.com/) - [Automatic Blog - Server Centric Development](https://automatic.com/blog/) - [OOPSpam Annual Report](https://oopspam.com/) ## 📝 Summary - Weekly WordPress discussion covering WordPress 7.0 beta release, AI developments, security updates, and industry trends - Panel discussion highlights tension between JavaScript frameworks and server-side development approaches - Key security updates include BitWarden audit response and password manager recommendations - AI discussion focuses on open source alternatives and cost-benefit analysis for WordPress users - WordPress 7.0 beta brings collaborative editing and design improvements ## 📖 Main Content This episode covers the evolving WordPress landscape with particular attention to the upcoming WordPress 7.0 release featuring collaborative editing capabilities. The panel discusses industry trends toward reconsidering JavaScript-heavy development approaches, with Automattic advocating for more server-centric solutions. Security remains a key focus with discussions about password managers and spam prevention. The conversation explores the tension between proprietary and open-source AI models, considering cost, privacy, and performance implications for WordPress users. ## 🗒️ Notes #### ⏱️ [00:00](https://youtu.be/VMzU4GhuuCI?t=0) — [H1] Introduction and Panel Setup - Key idea: Weekly WordPress roundup with international panel - Evidence: "This Week in WordPress #367, Monday February 23rd" #### ⏱️ [02:30](https://youtu.be/VMzU4GhuuCI?t=150) — [H2] Weather Check-in and Community Building - Key idea: Community engagement through casual conversation - Evidence: "Minus 5 here... visibility nil, blizzard warning" #### ⏱️ [05:45](https://youtu.be/VMzU4GhuuCI?t=345) — [H1] Andy Bell's AI Principles Discussion - Key idea: Some agencies taking principled stance against AI usage - Evidence: "He refuses to work with AI companies and won't use AI tools" #### ⏱️ [08:20](https://youtu.be/VMzU4GhuuCI?t=500) — [H2] MainWP's 12-Year Anniversary Review - Key idea: Long-term WordPress business sustainability and team motivation - Evidence: "12 years in the WordPress space is quite a marker" #### ⏱️ [12:15](https://youtu.be/VMzU4GhuuCI?t=735) — [H1] Automattic's JavaScript Framework Critique - Key idea: WordPress ecosystem questioning JavaScript-heavy development - Evidence: "A modest server-centric development proposal" #### ⏱️ [18:40](https://youtu.be/VMzU4GhuuCI?t=1120) — [H2] BitWarden Security Audit Response - Key idea: Transparent handling of security research findings - Evidence: "Security through Transparency - ETH Zurich audits" #### ⏱️ [22:30](https://youtu.be/VMzU4GhuuCI?t=1350) — [H3] Open Source Text-to-Speech Models - Key idea: Lightweight AI models for WordPress accessibility features - Evidence: "Kitten TTS is just 25 megabytes with 8 voices" #### ⏱️ [26:45](https://youtu.be/VMzU4GhuuCI?t=1605) — [H2] Spam Attack Trends and Residential IPs - Key idea: Increasing attacks from compromised consumer devices - Evidence: "Attacks coming from residential IPs, smart devices compromised" #### ⏱️ [30:20](https://youtu.be/VMzU4GhuuCI?t=1820) — [H1] Open Source AI Models vs Proprietary - Key idea: Quality gap closing but adoption challenges remain - Evidence: "90% quality difference closed within 13 weeks" #### ⏱️ [35:50](https://youtu.be/VMzU4GhuuCI?t=2150) — [H2] Hardware Requirements for Local AI - Key idea: Performance tradeoffs between local and cloud AI - Evidence: "Takes 45-50 seconds locally vs 5 seconds with ChatGPT" #### ⏱️ [42:10](https://youtu.be/VMzU4GhuuCI?t=2530) — [H3] Internet Archive and AI Scraping Conflicts - Key idea: AI scraping threatens historical web preservation - Evidence: "Guardian, NY Times, Reddit blocking Internet Archive access" #### ⏱️ [45:30](https://youtu.be/VMzU4GhuuCI?t=2730) — [H1] WordPress 7.0 Beta 1 Release - Key idea: Collaborative editing features approaching production readiness - Evidence: "Collaborative editing which many thought would be a roadblock is coming" ## 📚 References - [This Week in WordPress Show](https://wpbuilds.com/live) - [WP Build Show Feeling Insecure](https://wpbuilds.com/) - [Kitten TTS GitHub](https://github.com/kittenml/kitten-tts) - [Open Router](https://openrouter.ai/) - [TechDirt - Preserving the Web](https://techdirt.com/) ## 🗣️ Paraphrase - WordPress 7.0 beta introduces collaborative editing capabilities that were previously considered technically challenging to implement - The WordPress ecosystem is reconsidering JavaScript-heavy development approaches in favor of server-side solutions for better performance - Security remains paramount with password managers being essential tools, despite recent audit findings being largely theoretical ## 📰 Newsletter This week brings exciting developments in WordPress core with the 7.0 beta showcasing collaborative editing features that promise to transform content creation workflows. The community is also engaging in important conversations about development approaches, with Automattic advocating for more server-centric solutions over JavaScript frameworks. For security-conscious users, the BitWarden audit response demonstrates the importance of transparency in security tools. Meanwhile, open source AI models are becoming viable alternatives for accessibility features like text-to-speech. Resources: [WordPress Beta Testing](https://wordpress.org/download/beta/), [MainWP Community](https://mainwp.com/), [Security Best Practices](https://developer.wordpress.org/security/) ## ✅ Action Items #### WordPress - Core: Test WordPress 7.0 beta collaborative editing features in staging environments - Core Performance: Monitor impact of collaborative editing on server resources and database performance - Accessibility: Evaluate open source text-to-speech models for improved content accessibility - Security: Update security recommendations to include password manager guidance - Documentation: Create guides for server-centric development approaches vs JavaScript frameworks #### Community - Contributors: Share experiences with WordPress 7.0 beta testing and collaborative editing workflows - Maintainers: Review JavaScript dependencies and consider server-side alternatives where appropriate - Sponsors: Support open source AI initiatives that benefit WordPress ecosystem #### OSPO - Security Management: Review implications of open source AI models for enterprise WordPress deployments - License Compliance: Assess licensing implications of various AI model integrations - Community Engagement: Foster discussions about sustainable development approaches ## 🧪 Hypotheses - WordPress 7.0's collaborative editing will drive increased adoption among content teams - Server-centric development approaches will gain traction as JavaScript fatigue increases - Open source AI models will become standard for WordPress accessibility features - Password manager adoption will accelerate following security audit transparency ### 📊 Methodology - Beta testing WordPress 7.0 collaborative editing across different hosting environments - Performance benchmarking of server-side vs JavaScript-heavy implementations - User experience studies of open source vs proprietary AI integrations ### ❓ Questions - How will collaborative editing impact WordPress hosting requirements and costs? - Will the server-centric development trend influence WordPress core development priorities? - What role should open source AI models play in WordPress accessibility standards? ### 🔮 Future Research - Long-term performance implications of collaborative editing at scale - Security implications of self-hosted AI models in WordPress environments - Community adoption patterns for server-centric development approaches ### 💡 Implications - Practice: WordPress developers may need to rebalance JavaScript vs server-side skills - Policy: AI model choices increasingly impact privacy and cost considerations - Security: Transparent security practices become competitive advantages - Education: Training needed for collaborative editing workflows and server-centric development - Funding: Open source AI models require sustainable funding mechanisms ### 🧠 Learning Styles Assessment Reading level estimate: Grade 12 #### 👀 Visual learners - Present: Live video format with multiple speakers and screen sharing - Improve: Code examples and architectural diagrams for technical concepts #### 🔊 Aural learners - Present: Podcast format with conversational discussion - Improve: Audio-only version with clear speaker identification #### 🗣️ Verbal learners - Present: Technical terms explained in context, analogies used - Improve: Glossary of WordPress-specific terminology #### 🧩 Kinesthetic learners - Present: Encouragement to test beta versions and experiment - Improve: Step-by-step setup guides for featured tools #### 🧠 Logical learners - Present: Performance comparisons and cost-benefit analysis - Improve: Decision trees for choosing between alternatives #### 🧑‍🤝‍🧑 Social learners - Present: Live chat interaction and community engagement - Improve: Discussion forums for continued conversation #### 🧘‍♀️ Solitary learners - Present: Podcast format allows for individual consumption - Improve: Self-assessment checklists for covered topics #### 🌿 Naturalistic learners - Present: Organic conversation flow and real-world examples - Improve: Ecosystem thinking about WordPress sustainability ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Video has spoken content suitable for audio-only consumption - Improve: Formal transcript with speaker identification and timestamps #### Keyboard navigation - Present: Standard YouTube player controls - Improve: Chapter markers for easy navigation between topics #### Contrast and readability - Present: Video player meets standard contrast requirements - Improve: High-contrast mode for live chat interface #### Text resizing and reflow - Present: Responsive video player - Improve: Adjustable text size for live comments #### Media and captions - Present: Live spoken content - Improve: Real-time captions and downloadable transcript #### Forms and interactions - Present: Comment forms with appropriate labels - Improve: Better error handling for failed comment submissions ### 🌍 DEIB Assessment #### Cultural breadth - Present: International panel with different perspectives - Improve: More diverse geographic representation #### Inclusive language - Present: Technical terms explained in accessible ways - Improve: Avoid cultural assumptions about weather and time zones #### Neurodiversity support - Present: Multiple format options (video, audio, text) - Improve: Chapter markers and topic timestamps #### Representation - Present: Technical experts from different backgrounds - Improve: More diverse speaker demographics #### Psychological safety - Present: Respectful discussion of different viewpoints - Improve: Clear community guidelines for live chat ### ✅ Logical Validity #### Strong reasoning - Deductive strength: Technical analysis of WordPress features based on beta testing - Inductive strength: Trend analysis from multiple data points across ecosystem - Causal reasoning: Clear connections between development choices and performance outcomes - Analogical strength: Effective comparisons between AI model types and capabilities #### Fallacies - Appeal to authority: Some reliance on "experts say" without detailed evidence - False dichotomy: Occasional presentation of JavaScript vs server-side as either/or choice #### Tone of voice - Overall tone: Conversational and accessible while maintaining technical accuracy ### 🛡️ DARVO Analysis - Deny: No concerning patterns of denial - Attack: No attacks on critics or opposing viewpoints - Reverse victim/offender: No victim/offender reversals detected - Emotional cues and rhetoric: Generally balanced and factual discussion ## 🎯 Conclusion This episode effectively covers the current state of WordPress development with particular attention to upcoming collaborative editing features and evolving development philosophies. The discussion of open source AI alternatives and security best practices provides practical value for WordPress professionals navigating rapidly changing technological landscapes. ## 🧱 Refactor Appendix ### 📚 Literature Split Pad Purpose: Stage CEE units that become Literature notes. ### [H1] WordPress collaborative editing transforms content workflows > [!quote] This Week in WordPress #367 by Nathan Wrigley via WP Builds. Published 2025-02-24. > "Collaborative editing which many thought would be a roadblock is coming" WordPress 7.0 beta demonstrates technical feasibility of real-time collaborative editing features. [view highlight](https://youtu.be/VMzU4GhuuCI?t=2730#H1) #xref #articleseed ID:rw_wordpress_collab ### [H2] Open source AI models closing quality gap rapidly > [!quote] This Week in WordPress #367 > "90% quality difference closed within 13 weeks" Open source alternatives achieving near-parity with proprietary models in short timeframes. [view highlight](https://youtu.be/VMzU4GhuuCI?t=1820#H2) #xref ID:rw_ai_quality_gap --- ### 🔖 Atomic Split Pad (Zettel Candidates) #### [Z1] Server-side development reduces JavaScript complexity - Applicable to WordPress sites prioritizing performance over interactivity - Reference to Automattic blog post on server-centric approaches - Links: [[WordPress-Performance]] [[JavaScript-Frameworks]] #### [Z2] Password manager adoption essential for security - Universal recommendation regardless of specific tool choice - Transparency in security audits builds user confidence - Links: [[WordPress-Security]] [[BitWarden-Audit]] --- ### ✏️ Fleeting Queue - [F1] Test WordPress 7.0 beta collaborative features — staging environment setup #fleeting #inbox - [F2] Research Kitten TTS integration for accessibility — 25MB model evaluation #fleeting #inbox - [F3] Review MainWP anniversary insights — 12-year sustainability factors #fleeting #inbox --- ## Full Transcript [02:00] Hello, good morning, good afternoon, [02:02] good evening wherever you are in the [02:04] world. And I can see some of you from [02:06] different parts of the world. It's very [02:07] nice. Lots of comments dropping in [02:09] largely about the weather which is [02:10] always interesting. Um, so let's just [02:13] sort of say what we're doing this week. [02:15] It is this week in WordPress number 367. [02:19] I feel that's the wrong number, but [02:21] let's go with that. That's what I wrote [02:22] down. 367. [02:24] It is Monday, the 23rd of February. [02:27] We'll be repackaging this as an episode. [02:28] It will come out into your podcast [02:30] player of choice tomorrow. Tomorrow [02:32] morning at the the highly unusual 7 a.m. [02:36] in the UK. I don't know why I picked [02:38] that, but I picked it and thought I'm [02:40] sticking to it. I'm being consistent. [02:41] So, it's gone out at 7:00 in the morning [02:44] in the UK for about eight years or so. [02:47] No reason, but at least it's there when [02:49] you wake up, depending on where you are. [02:51] The idea is to talk about WordPress. Um, [02:54] but increasingly we end up talking about [02:56] all sorts of different things. And I [02:57] think there'll be quite a lot of [02:58] security stuff, possibly AI. Well, I say [03:01] possibly AI stuff. It's mostly AI stuff [03:04] these days. There's there's no oxygen [03:07] left in the room. Um, but in order to [03:09] make that happen, I'm always joined by a [03:11] panel of um, experts, let's put it that [03:14] way. And, um, today I'm joined by these [03:17] three fine people. There, there, and [03:19] there. And first one over here is uh, [03:23] Sir Tim Nash. Hello Tim. How are you [03:25] doing? [03:26] >> Hello. I I just can't remember when I [03:28] got kned. You'd think it would be an [03:30] important thing in my [03:31] >> Yeah, honestly, this whole world kne [03:34] down guy with a sword coming at me, but [03:36] no. Budget cuts. Budget cuts. [03:38] >> Yeah. Yeah. Yeah. Yeah. The royal family [03:41] are having a tough time. [03:43] >> Yeah. [03:44] >> Wonder if they'll start charging. [03:46] Anyway, let's not get into that. That's [03:48] definitely off topic. Um, but how you [03:51] doing? You all right, Tim? [03:53] >> Ah, yeah. I'm here. [03:56] That's about as much enthusiasm as you [03:57] can reach the moment. [03:59] >> That's the level of enthusiasm we're [04:01] gonna get. I ask the the contributors to [04:04] give us a little potted bio. And so Tims [04:06] Tims has Tims has always excited me in [04:09] the past. Uh I never read it just in [04:11] case it's something that requires a bit [04:13] of inspiration at the start. So here it [04:14] comes. I genuinely haven't read it. [04:16] Right. Here it is. Tim scares people and [04:19] oh no. Tim scares people and no more so [04:22] than Nathan Wrigley for he fears the [04:24] intros. He will be sing will he be [04:26] singing, dancing an intro or just saying [04:29] hello? I don't know. Uh today though he [04:32] should not fear he should fear not as [04:33] the no intro is oh is this intro is [04:36] really tame. Thank the lord. Tim is a [04:39] professional doom speaker and security [04:42] consultant specializing in WordPress. [04:45] You can find about you can find out more [04:46] about him at timnash.co.uk [04:49] or listen to his semi-regular chat with [04:52] Nathan on the WP Build Show Feeling [04:54] Insecure. We should do that again. We we [04:57] should we haven't for a recording for a [04:59] long time a long time and uh yeah let's [05:01] get back on the wagon with that but I I [05:03] am so pleased that there's there was [05:06] nothing that I had to sing or do as a [05:08] sea shanty or any of that this week this [05:11] time. [05:12] >> Don't worry I'm back in a few weeks [05:14] time. [05:14] >> Yeah. Okay. All right. I'll hold my [05:17] breath. Uh but thank you for joining us. [05:19] I'm also joined by Mark Benzane. [05:22] Hello Mark. [05:23] >> Hello. [05:24] >> How you doing? It's always like a [05:26] pleasure to drag you out of your bed at [05:28] 3:00 in the morning or whatever it is. [05:30] >> It's it's it's the pleasure is all mine. [05:33] And I mean that 100% sincerely. [05:36] >> What time is it? [05:37] >> Uh right now it's actually uh it's at [05:41] six o'clock. But by the time I roll out [05:44] of bed and, you know, take my shower and [05:47] do all the things and make sure that my [05:49] setup is right, especially this weekend, [05:51] my my computer crapped out and I had to [05:53] totally wipe and reload everything. And [05:56] I'm like, [05:57] >> yeah. So, so I'm like scrambling around [05:59] all weekend trying to make sure that [06:01] everything will work for the show today [06:02] and cuz I'm not going to have an hour [06:05] ahead of time to make sure things work. [06:07] >> No, no, no, no. That would mean getting [06:09] up at 2:00 in the morning. Well, that's [06:11] what that would mean. Exactly. Yeah. [06:13] >> If you're watching the show, you too [06:14] will notice that Mark spends his entire [06:17] life inside of a box. [06:18] >> I did. [06:19] >> Black box that follows him around [06:20] wherever he goes. [06:21] >> But I think outside of it. I think [06:23] outside of my my physical self is inside [06:27] the box. [06:28] >> You picked up what I was putting down. [06:29] Well done, Mark. That was great. Uh [06:31] Mark's bio goes like this. Uh Mark is [06:34] the partnerships and community lead at [06:36] MainWP. And though a totally contrived, [06:41] sorry, and through a totally contrived [06:43] coincidence, the timing lines up with [06:45] MainWP's 12th anniversary. Oh, nice. [06:48] Nice. Birthday celebrations. 12 years in [06:50] the WordPress space is quite a marker [06:52] and it felt right to acknowledge it [06:54] whilst he's here with us. Mark has been [06:56] part of the ecosystem for ages, working [06:58] closely with agencies, developers, and [07:00] product teams who manage sites day in [07:02] day out. His work at MainWP centers on [07:04] partnerships, community and helping [07:06] shape tools that support people building [07:09] and maintaining WordPress sites at [07:12] scale. Thank you, Mark. Really [07:14] appreciate it. [07:15] >> Thank you for having me. [07:16] >> Yeah, you're very welcome. And joining [07:18] us for the first time is Onar Aliy. [07:21] Hello. How you doing? Have I said you're [07:22] >> Thanks for having me. [07:24] >> Yeah, thanks for having me. I'm very [07:26] happy to be here. [07:27] >> Yeah, you're very welcome. Um, as as [07:30] Onar, it's his first appearance. there's [07:31] this empty bio box and no doubt on the [07:34] second appearance it will be full up [07:36] with junk, you know, because that's what [07:38] we do. We just mess it up. But on this [07:39] first time around, it's very obvious [07:41] that it's going to be sensible. So, here [07:43] is uh Onar's very sensible bio. He's the [07:46] founder of OOP SPAM. I'm not entirely [07:49] sure how you pronounce that. He can [07:50] correct that for us in a moment. [07:51] Formerly cyber security researcher [07:53] focused on online abuse and spam. And [07:56] he's based in Minneapolis, Minnesota. [07:59] How do we say it, Onar? [08:02] when you say we say oops ma'am. [08:04] >> Okay. I did actually think that would be [08:06] what it was but I didn't want to Yeah. [08:08] Okay. [08:09] >> I started a long time ago and now I'm [08:12] sinking back. Probably wasn't the best [08:14] name but [08:15] >> Yeah. [08:16] >> Yeah. Well, as you're about to find out, [08:18] I called this thing WP Builds and we [08:20] basically talk about cats. We has spent [08:23] ages on an episode the other day talking [08:25] about alligators. So [08:26] >> Oh, that's [08:28] hold on. Uh we'll see how that goes. If [08:31] you are joining us and you want to make [08:33] the show a little bit more interesting [08:34] and lively um then please make some [08:37] comments. That would be absolutely [08:38] lovely. The best way to do that is to go [08:41] here. Um Mark is sporting the little [08:45] banner there. Go share this stream [08:47] wpbuilds.comlive. [08:50] If you go there, you've got um you got [08:52] like two options really. The the the [08:54] most common option I think is to hit the [08:56] box on the side if you're on a desktop. [08:58] It's Google. So, it's YouTube comments. [09:01] So, if you're logged into a Google [09:02] account somewhere, you can comment like [09:03] that. However, if you don't want to use [09:06] a Google account or you don't have a [09:07] Google account or you mistrust Google or [09:09] any other reason for that matter, then [09:12] there's a box inside the video player [09:14] itself top right. It says start chat or [09:18] live chat or something like that. Click [09:20] on that and you don't need to be logged [09:21] into anything. You can uh be entirely [09:23] anonymous. And so please share it with [09:26] your friends, relations, animals, [09:29] pets, poltergeists, whatever the heck [09:31] you've got going on over there. And it [09:33] looks like some people have jumped in. [09:35] So here we go. We often get the weather [09:37] forecast. [09:39] Oh dear. Uh, and in Boston, James Laauo [09:43] is joining us. And it's 30. Oh, that's [09:45] brutal. It's minus one degrees [09:46] centigrade. Birds, birds, animals, tree, [09:49] branches, people fly at a 90 degrees [09:52] angle. Visibility nil. Officials, [09:54] blizzards warning roulette remain in [09:56] place until 7:00 a.m. ESC. Good. Good [09:59] morning. Oh, that sounds bad. [10:01] >> Well, it's minus5 here. [10:03] >> Oh. Oh. Oh, that's not real. That's like [10:06] the temperature on Mars or something. [10:09] >> But at least the internet works. So, [10:12] >> electrons love it. They love all this [10:15] cold. It's perfectly. [10:17] >> Uh Courtney Robertson's joining us, but [10:19] she's saying hello to James. Uh, she's [10:22] given a comment about the weather there. [10:23] Just two inches here. I guess that's [10:25] snow she's talking about. Oh, David [10:28] Warmsley, my partner in crime, David [10:30] Warsley. He says, "Hello, first time [10:32] catching the beginning of this show. [10:34] What a birthday treat." I'm presuming [10:36] it's your birthday, David. And if it is [10:38] your birthday, happy 92nd birthday. It [10:42] must be that at least by now. Uh, [10:45] something along those lines. I don't [10:47] know how old David is. I think he's [10:48] older than me then. But happy birthday, [10:51] David, if it is your birthday. Uh, the [10:53] webinar is also joining us. Hello, [10:55] beautiful people. How you doing? And [10:57] gosh, there's quite a few, isn't there? [10:58] Lots of people saying hello to each [11:00] other. Oh, come on. Rob rub it in, won't [11:03] you? David's in Goa where it's 27 [11:08] 27° centigrade. All I can say is the [11:11] electrons hate it. [11:13] >> I imagine the humidity is horrible [11:15] though. [11:16] >> Yeah. Yeah. [11:16] >> Go on. Go on. Tell me the humidity. [11:18] >> It's his birthday. He's getting old and [11:20] the humidity is horrible. That's it. [11:22] That's how we're going to how we're [11:23] going to reconcile it. Uh Ross Win, [11:27] hello. He says he's got the day off so I [11:29] can be live here today. A you are [11:32] friends, not work. That's so nice. Thank [11:34] you for joining us, Ross. Appreciate it. [11:36] Uh 12 degrees centigrade here in the UK [11:38] and very British. Yeah, I got the same [11:41] 12 degrees centigrade and just greyness. [11:44] You know, it's what we get every single [11:45] day of the year. Any other comments [11:47] joining in? Who we got here? Marcus [11:49] Bernette. Sorry. He says, "I had to [11:51] cancel you, Nathan." Oh, yeah. [11:55] Luckily, Onar's here. You know, that's [11:57] all I have to say about that. Marcus [11:59] wasn't able to come because he had other [12:01] things. And so, Onar very graciously [12:03] jumped in on Friday. And uh that's [12:05] amazing. Um testing WPB to one. We'll [12:08] get to that in a moment. As Jabber, [12:14] he's on skating on thin ice, is he [12:16] James? [12:18] Do I jabber? I think I jabber. [12:20] >> It's cold there. It's very thick ice. [12:22] Trust me. [12:23] >> Yeah, that's right. I think uh I think [12:26] is there any other comments that anybody [12:27] can say that I ought to put on the [12:29] screen because I'm kind of lost in them [12:30] now. There's too many to come up. [12:32] Anybody quickly? [12:33] >> Well, Courtney Courtney asked, but I [12:35] said I responded already, but uh [12:37] Courtney asked if the reset of my [12:39] computer means that the lights will not [12:41] go out. Uh I cannot guarantee anything. [12:44] Um, [12:45] >> Mark Mark's in a black box already, but [12:48] at some point the lights are go out, so [12:49] it'll be entirely black. It's really [12:51] funny when it happens. If you sit really [12:54] still, Mark, for like five minutes and [12:56] say nothing, we can all enjoy your [12:57] lights going out. Uh, okay. So, anyway, [13:00] there we go. That's the long [13:01] introduction. Really appreciate you all [13:03] being here. Keep dropping in some [13:04] comments. So, so, so the the idea is to [13:06] talk about WordPress. This is us, by the [13:08] way. Bit of short short promotional [13:10] stuff at the beginning. If you like what [13:11] we're doing over here, put your name [13:12] into that box. This is wpbuilds.com. [13:15] No, don't put your name. Put your email [13:17] address into that box. We don't need [13:18] your name. Just want your email address. [13:20] And then we'll send you a couple of [13:22] emails each week. One when we put this [13:24] out as a podcast. That'll happen [13:25] tomorrow when I've edited this. Um, and [13:28] then we'll put an episode out every [13:30] Thursday as well, and I'll send you an [13:32] email about that. And this week, uh, I [13:35] was lucky enough to have the fabulous [13:38] Andy Bell. Uh, Andy Bell. I think in [13:41] years gone by Andy Bell was more of a [13:44] WordPresser than he is now. I think it's [13:46] been many years I think since he's used [13:48] WordPress uh in in anger if you know [13:50] what I mean. But I had a chat with uh [13:52] Andy because I saw him at a conference [13:54] last year which Tim Tim was there as [13:56] well. It was in London last year and uh [13:58] Andy was on stage and he wasn't really [14:00] talking about WordPress. He was talking [14:02] about sort of agency life and things [14:04] like that. But the one of the things [14:06] that caught my attention was his [14:08] position on AI. Now Andy is very [14:10] principled about AI and he's very very [14:13] forthright. He doesn't want anything to [14:15] do with it. He doesn't want anything he [14:17] doesn't want his agency touching it. So [14:19] the content that they produce, the code [14:21] that they write, the websites that they [14:23] launch is all humanmade. [14:27] As soon as those words came out of my [14:29] mouth, I realized the error. Uh it was [14:31] all made by his agency, not humanmade. [14:35] Forgive me. Um [14:37] but uh but um anyway but he also won't [14:41] won't work with AI companies. So if an [14:43] AI company comes to his agency and says [14:45] will you build us a website? The answer [14:47] is going to be no. So we explore his [14:49] fairly principled position here. And I [14:51] got to say the more he talks the more I [14:55] was nodding my head you know the [14:57] dominoes kind of started to fall for me [14:59] and uh he was able to encapsulate it [15:01] really well. So I don't suppose any of [15:03] you three have got anything to add about [15:04] that. This is my promo bit at the [15:06] beginning, but I've dropped the subject [15:07] of AI and his principle on it. [15:10] >> I feel on principle, I should say. Did [15:12] the word lite turn up in the [15:14] conversation? [15:14] >> Uh, it should have done, shouldn't it? [15:16] Yeah, probably. But I I know I know what [15:19] you mean. [15:20] >> I think his is not really captured [15:22] around the technology. He's obviously [15:24] got a notion that he can do whatever the [15:26] heck an a an AI can do when it comes to [15:29] design. He's got the skills. He doesn't [15:31] need the shortcuts. And so whilst it [15:34] might save him a bit of time and [15:36] probably some money, go just go and have [15:38] a listen to him. I I know what you mean [15:40] though. But um I found myself nodding is [15:43] all I can say. [15:44] >> Uh anybody else, Mark Onar, anything to [15:47] add about that before we move on. [15:49] >> I I respect anybody who has principles [15:51] and sticks to them. Yeah, [15:53] >> that's that's you know um and so I I [15:57] have a great amount of respect for what [15:58] he's saying and if he can and I think [16:00] it's always important to be able to do [16:02] what the AI is doing. You have to be [16:04] able to understand what it's doing so [16:07] that you can either fix it or or or not. [16:10] So if you choose not to do it, it means [16:12] you know what you're doing. Well, and [16:14] the the other interesting bit was that [16:16] as as the sort of the conversation um [16:19] developed, not only was I kind of [16:21] getting more and more into it, but I [16:23] could see where he was coming from. And [16:25] the bit that kind of made the principles [16:27] stick for me was the fact that it really [16:30] hurt his business in the year 2025. [16:34] Incredibly successful and it had hurt, [16:37] you know, that having this principle had [16:39] hurt his business without a doubt. He [16:41] was very open about that. But what he [16:44] had not done publicly is he hadn't said [16:47] out loud the bit about the AI. He' just [16:51] not taken work that required him to use [16:54] AI. He hadn't taken work from AI [16:57] companies and he hadn't used AI. So [17:00] anyway, the knock-on effect was as soon [17:02] as he started saying the quiet thing out [17:05] loud, I refuse to work with AI, suddenly [17:09] his roster got filled up again. So it [17:11] turned out that for a small subset of [17:14] people and he presumably is an early [17:16] mover that that rounded off his 2025 [17:20] really well. Now we'll have to see how [17:21] long that continues, but it turned out [17:23] that that principle resonated with a lot [17:25] of people who needed a website. Um but [17:29] we'll have to wait and see how that that [17:31] works over time. Okay. Anyway, there we [17:33] go. Anybody else? Before I move on, [17:36] please keep your eye on the comments, [17:37] you three, because um if there's [17:39] anything that crops up, there's too many [17:40] at the minute for me to keep my um my my [17:43] eyeballs on. Uh by the way, um he has a [17:47] u like a course platform called uh [17:50] Pikilly. It's p i ca l i l. Pilly with [17:55] the li after the dot. And um a lot of [17:59] the courses are about this kind of [18:00] thing, you know. So mindful design is [18:02] one of them. I just thought I'd mention [18:04] that in case anybody was keen to pick up [18:06] on what it is that he's doing. Go and [18:07] check it out. Right. Okay. So, moving on [18:10] to user submitted content or in this [18:12] case guest submitted content. Here we [18:13] go. MainWP. You brought this to our [18:16] attention. It looks like the year in [18:17] review. How did MainWP go, Mark? [18:21] >> 25. [18:22] >> Well, first of all, Dennis asked me to [18:24] write the year in review this year, [18:26] which is actually a pretty intimidating [18:28] task, I got to say. And and I was kind [18:30] of like when you go back and and look at [18:32] everything that happened, it kind of [18:34] blows your mind at blew my mind as to [18:37] how much uh the team just did last year. [18:41] And I I think about it once again in the [18:45] frame of this 12-year anniversary thing. [18:47] And I just am blown away constantly that [18:51] they are so motivated every single day [18:55] to just keep pushing and pushing and [18:57] pushing after 12 years. I mean, that [18:59] just is astounding. So, that's the big [19:02] takeaway for me from this. The rest of [19:04] it goes into details as to how many [19:06] things we pushed out and and where we're [19:09] going and and all of that. Uh, and I [19:11] won't get into all of that. I I would [19:13] appreciate it if people would read it. I [19:14] appreciate you letting me share this uh [19:17] with the audience and I hope people read [19:19] it. And uh and uh I don't want to take [19:22] up too much more time other than that, [19:24] but uh I just I I just can't talk about [19:29] the group of people that Dennis has put [19:31] together at MainWP. It's just amazing [19:34] group. [19:35] >> Nice. Nice. Okay. Well, probably a quick [19:38] Google. Everything that we mention in [19:40] this show will be in the show notes [19:42] tomorrow. So when the podcast audio [19:43] drops, there'll be a link to a blog post [19:46] which is basically a bunch of links. [19:48] Some of them are in the show. There's [19:49] ones that didn't make it into the show, [19:51] but all of the ones that we're going to [19:52] mention will be in there. So, this will [19:54] make it uh into that list. It'll [19:56] probably be under the community section [19:58] if you go there, or possibly the plug-in [20:00] section, depending where I [20:01] >> where I drop it. But go and have a look. [20:02] Uh or you could simply Google at this [20:04] moment, couldn't you? Year of momentum. [20:06] A year of momentum 2025 year in review. [20:09] Nice that you love your team as much as [20:11] that. That's brilliant. [20:12] >> I I feel so lucky. Honestly, I I I went [20:15] from being a, you know, a an owner of a [20:18] company and a founder of a company to [20:20] being in a company where I feel like I'm [20:22] a founder even though I'm not. So, it's [20:25] it's it's really I I feel very fortunate [20:28] to have landed where I did. [20:29] >> Yeah. Oh, that's lovely. I'm just going [20:31] to pause the entire show very quickly [20:33] because I just want to say [20:36] over a period of about three months now [20:38] this person has dropped into the show um [20:42] Gank and every single episode I have no [20:47] idea what his name is and every episode [20:49] he says what his name is and I say I'll [20:51] remember it for next time and I never [20:53] remember it. So I've actually written it [20:55] down on a piece of paper hoping they [20:57] drop in today. I think you should name [21:00] the show after him. Yes. Just so that [21:02] you have an episode that's like [21:04] >> Yeah. Well, he's joining us from [21:05] Amsterdam. And of course, thank you, [21:06] Brian, for uh Oh, see what I did there? [21:09] Uh, it's Gerard. Thank you, Gerard, for [21:13] joining us. I really appreciate it. And [21:14] I I appreciate me. Sorry, I don't [21:17] appreciate me. I appreciate the fact [21:18] that you're coping with the fact that I [21:20] relentlessly, you know, been unable to [21:22] remember who you are. I really [21:23] appreciate it. And um, nice mention for [21:26] you, Tim. I think it was Ross I saw in [21:28] the comments who said, "It's lovely to [21:29] see you." Uh, what did he say? He said, [21:32] "Oh, no, it was Reese W. Nice to see Tim [21:35] again." That's [21:36] >> a Thank you. [21:37] >> Yeah, that's right. The bearded club. [21:40] They you you stayed together, you guys. [21:44] Okay. Have we got any nice comments in [21:46] there that Oh, that's very nice. Ross, [21:48] he's sort of saying, "Go and listen to [21:49] the interview, I presume, about Andy." [21:51] Uh, and staying on that topic, he says [21:54] Andy also said, so this is the previous [21:55] thing with Andy Bell. Um, that he'd got [21:58] some business from a client that had [21:59] been burned. Oh, that's interesting. [22:01] Yeah. By an agency that had used AI. Um, [22:05] and we're looking for somebody that [22:06] explicitly did not, which was kind of [22:09] interesting. [22:12] >> Um, and then David Wsley joins in. [22:14] Primary Andy problem solves and writes [22:17] HTML and CSS better than LLM seems to be [22:20] able to manage. I don't think you could [22:21] deny that somebody of his capacity [22:24] that's very likely to be true, isn't it? [22:26] Okay. All right. Let's move on a little [22:29] bit and we'll go here. So, that was the [22:31] bit that Mark dropped. Now, we've got a [22:32] couple coming from Tim's side of things. [22:34] I didn't get a chance to read this one, [22:36] Tim. Sorry about that. Um, so tell us [22:39] why this one's in. It's from Automatic. [22:41] >> So, it's sort of slightly slightly links [22:45] into Andy. Uh so this is a uh blog post [22:49] by somebody from automatic. They're on [22:52] the performance team at Automatic. Not [22:55] it's not specifically WordPress related [22:58] directly, but um it's basically a blog [23:02] post arguing we use too much JavaScript. [23:06] >> We use too React is React is not a [23:10] library. It's a framework. We use these [23:12] giant frameworks. They're ridiculously [23:15] complicated. They are really bad for [23:19] performance. They very rarely do the [23:22] things that we actually want them to do [23:23] well and we keep use trying to use [23:26] JavaScript when actually we have this [23:28] perfectly good servers behind the scenes [23:30] that are full of processing power that [23:32] could be doing the job for us. that a [23:35] lot of the time we struggle with [23:37] performance on websites is because not [23:40] necessarily the servers slowing down, [23:43] but your client has got to make a [23:45] 100,000 different requests backwards and [23:47] forwards to the server via the [23:49] JavaScript Ajax requests and JavaScript [23:52] is pulling and doing a lot of issues [23:55] when it comes to performance. And it's [23:57] basically a little bit of a h maybe we [24:01] should perhaps not use as much [24:04] JavaScript or only use JavaScript when [24:07] we need to and use it for what it's [24:09] really good at. Um, which I found quite [24:13] interesting. So it's not a get rid of [24:15] Ajax, get rid of React, get rid of all [24:18] of this. It was just a maybe we should [24:21] rethink that your if you're jumping into [24:24] some a project and your first reaction [24:26] is to jump and get react maybe consider [24:29] if that's the right course. And I found [24:32] that fascinating partially because it [24:34] very much mirrors my own viewpoint. So [24:37] of course I'm biased towards this as I [24:39] like this makes total sense. Uh I think [24:41] there's a lot of people in the comments [24:43] who will find this approach like oh yeah [24:46] that's what I do. that should be what we [24:48] should do. [24:49] >> But also who it came from. [24:51] >> Yeah, [24:52] >> it's on the automatic blog, [24:54] >> right? [24:55] >> I I'd like to think that this was a [24:58] little bit of a a flag has gone up, but [25:00] maybe we have gone too far the wrong [25:02] way. Um, but we'll see. But it's a it [25:06] was it's a really good read. It's a [25:07] really wellthoughtout read that um for [25:11] me uh reflects a lot of my own [25:14] viewpoints. So, I was sort of I was I I [25:17] very much was reading this and nodding [25:20] my way along. So, I think [25:23] >> I can more or less guarantee that David [25:24] Warsley in the comments will be warming [25:26] his hands on on this. He shares the [25:30] exact same opinion as you. Uh, you know, [25:32] buy you something when it's not [25:34] necessary. But you're right, that was [25:36] the bit that I took. I read the the [25:38] title and I read that it had come from [25:39] automatic and then I got to the bit [25:42] where it was obviously taking its first [25:44] like little targeted aim toward React in [25:48] particular [25:49] >> and I did think this is curious. [25:51] >> It is he he does take quite a strong aim [25:54] at React but then does emphasize that [25:56] it's not specifically React. It's more [25:58] that React is [26:00] >> act and I think his argument basically [26:02] is you never load just one JavaScript [26:04] library. [26:06] It's part of the uh philosophy of [26:10] JavaScript development seems to be well [26:12] I can just reach for this and I can [26:13] reach for that [26:14] >> and so you do end up with behemoths very [26:17] quickly and without even realizing it [26:19] you're you're suddenly you've got huge [26:21] amount of dependencies there that you [26:23] weren't expecting and yeah it's as I say [26:25] it's a really good read uh even if you [26:28] take away from it that is like that's a [26:30] bit extreme I if it just gives you pause [26:32] for four that maybe I don't necessarily [26:35] really need to bring in that full [26:36] framework to do this oneline thing. Or [26:39] maybe it could be just that you take the [26:41] opinion, oh actually I don't need to use [26:44] I can rethink how I'm using my [26:46] JavaScript and make sure that I'm first [26:49] of all the server can do this on its [26:51] own. you're almost certainly getting [26:53] going to have a better experience in [26:54] terms of accessibility, probably a [26:56] better experience in terms of access um [26:59] things like SEO and almost certainly a [27:01] better experience from a security [27:02] perspective as well because if you're [27:04] simplifying things generally that makes [27:06] it more secure, not always, but [27:08] generally the bit that's kind of [27:11] interesting to me is I would imagine [27:13] that the automatic blog has some level [27:15] of editorial control. I imagine that you [27:18] know uh whoever wrote this article and [27:21] the name has now gone Sergio Gomez uh is [27:24] the automatician who wrote this. I [27:26] presume there's some level of [27:28] inspection. So it is it is kind of [27:30] interesting to me that this has been you [27:32] know this has been given the the green [27:34] light and I I wonder I just wonder if [27:37] this is a first [27:40] in a cascade of similar things in the [27:43] year 2026. I think maybe that's what [27:46] you're I don't know. Maybe if you're [27:47] wondering that as well, Tim. [27:49] >> Yes. Not sure. [27:50] >> Well, no. Hoping maybe. I know. I feel [27:53] like this could either be a one shot and [27:54] it's like, "Oh, that's nice." Or we [27:56] could see this happening more and more. [27:59] >> And there is a little bit of a case of [28:02] >> as we are getting more and more AI slop [28:06] and a lot of if you ask a AI to build [28:10] something, it will pull in all of these [28:12] libraries and things. I can see the the [28:15] Andes of this world and people going [28:18] actually I can [28:20] win this battle by doing less and [28:24] >> producing good quality simple solutions [28:27] rather than pulling in the 50 million [28:29] libraries. So I can see it being used [28:31] that way as well. So [28:33] >> thank you. Okay, that's really [28:34] interesting. Uh I didn't say the title [28:36] of the piece and I have I'll read it in [28:38] because if anybody's uh listening to [28:40] this primarily this is on the [28:41] automatic.com. So with 2T's blog and [28:44] it's called a modest server centric [28:47] development proposal. So you can go and [28:50] look for that. It was written on the [28:52] 17th of Feb this year. So go and check [28:55] that out. Right. And there was another [28:56] one coming from you. Sorry I apologize. [28:58] I didn't let Mark and uh Onar have their [29:00] go on this if they wish. Otherwise we'll [29:02] move on. [29:04] It's [29:04] >> above my pay grade. [29:06] >> Okay. Me too. [29:08] >> I think you can find markets below your [29:10] pay grade. [29:11] >> Oh is that what you think? Okay. Okay. [29:13] >> Uh, okay. Right. Here we go. Um, next [29:16] one then again from Tim. Um, Tim's thing [29:19] is obviously security. I use Bit Warden. [29:21] I I don't know if they have products [29:23] outside of the sort of password manager, [29:24] but that's the length and the depths [29:26] that I have from them. And obviously it [29:29] being a password manager, I have this [29:31] assumption that it's going to be secure. [29:34] So, what's this one? What are we looking [29:36] at here? So this is a a blog post from [29:39] Bit Warden themselves called Security [29:40] through Transparency. ETHZurich audits [29:43] Bit Warden and cryptography against [29:45] malicious server scenarios. uh ETH [29:49] Zurich actually audited three or four [29:51] different password managers and claimed [29:53] that there were some serious faults and [29:56] I got pinged a lot [30:00] like every friend I know who I said go [30:04] use this go use a password manager [30:06] pretty much together it's like they got [30:09] sent the same email sent me going oh my [30:12] god it's all really had bad and it's [30:13] insecure it's like no no no the the the [30:17] group managed to find if you compromise [30:20] the bit warden servers and the clients [30:23] in this scenario and this happens then [30:26] something bad might happen. [30:28] >> Oh okay that's quite a lot of this is [30:31] there were variations on the theme for [30:32] each one of the password managers. So um [30:35] the reason I chose this as a good uh [30:37] post is because I thought it was a [30:38] really well-handled response from Bit [30:41] Warden because they didn't say this is a [30:43] non-issue. I thought that was really [30:45] good of them because they could have [30:46] just come out of the gate going, "Uhhuh. [30:49] Uhhuh. No." But they didn't. They [30:51] actually said, "We've taken things from [30:53] this. We've learned things from the from [30:55] this paper. We agree it's hypothetical. [30:58] Um, but honestly, [31:02] let's go. You know, we we there isn't [31:05] actually a problem, but we've learned [31:06] stuff." But it also gets gives me the [31:08] opportunity to highlight that you should [31:10] be using a password manager. um that and [31:13] the one of the reasons that I wanted to [31:15] highlight Bit Warden is that Bit [31:16] Warden's um behind the scenes both the [31:20] client is open sourced and the um the [31:23] server infrastructure is open sourced [31:24] and you can run something like uh Vault [31:27] Warden on your own infrastructure as [31:29] well. So, [31:30] >> okay, [31:31] >> there if you were if you haven't yet [31:34] your password manager [31:36] >> and I'm I'm disappointed in you, but [31:39] there are options out there. Uh, [31:41] obviously, if you have a password [31:42] manager of your own choice, go with that [31:44] one. Please don't don't change because [31:46] of somebody saying that Bit Warden's [31:48] more interesting because it's [31:49] transparent because of open source. [31:51] >> But that's why I wanted to highlight [31:53] this. [31:53] >> Yeah. Thank you. [31:54] >> To put a lot of people to bed so they [31:55] can stop DMing me about [31:57] >> Yeah, stop stop DMing Tim. Um, but I'm [31:59] glad because this is the one that I [32:01] settled on a few I don't know two or [32:03] three years ago when Last Pass just sort [32:04] of seemed to misstep horribly, [32:06] especially from a PR point of view. Um, [32:09] I moved over and pushed everything over [32:11] to bit one. So, I'm happy um that you [32:14] have the uh you know that you got the [32:15] skinny on this. Uh, by the way, Tim [32:18] paused about about a minute ago for [32:20] about three seconds and he scowlled. [32:22] There was a moment where Tim Tim scout [32:24] it was a very effective scallally. He [32:25] was trying to encourage everybody to use [32:26] a password manager. So, unless you want [32:28] to be on the receiving end of Tim's [32:30] scowlling, then get yourself get [32:32] yourself a part of I feel there should [32:35] be scowling on your website, Tim. Uh, [32:38] for me, yeah, [32:39] >> I'm a nice person. I don't I I want to [32:41] have clients. [32:42] >> Okay. Scowlling at them before they sc [32:45] my clients. [32:45] >> I would go and get a password manager. [32:47] It's not worth it, everybody. The other [32:49] thing was Mark's lights went out. I [32:51] don't know if anybody noticed that, but [32:52] he scuttled away very quietly and got [32:55] his lights coming back on. But there we [32:57] go. Few comments in coming in. Um, [33:01] where was it again? So, going back to [33:03] the um the bit that we were talking [33:05] about a minute ago. So, Andy Bell and [33:07] what have you. Um, let me just put it [33:09] like that. David Warsy says, so in [33:12] regard to No, it wasn't. It was about [33:14] the automatic piece and React and what [33:16] have you. Um, David Warsy says, "I just [33:18] wish WordPress was more active in the [33:19] W3C rather than looking to the corporate [33:22] web all the time." Uh, there we go. And [33:26] uh apparently Tim, you're a sucker for a [33:28] clickbait title. [33:30] >> Yeah, I'm not quite sure which one. [33:32] >> Denny, I think Scowl. Tim. Scowl. Just [33:36] give him the Yeah. Yeah. David knows his [33:39] place now. He's No more comments from [33:41] you, David. I feel like the soup. What [33:44] was it? The soup thing in Seinfeld. No [33:46] soup for you. No comments for you, David [33:49] Wsley. You're you're in the doghouse. [33:50] Right. Moving on. uh the so there's [33:53] three that are coming up from ONAR and [33:55] the first one I don't know if the this [33:57] GitHub repo is something that you're [34:00] involved with personally but we're [34:02] looking at kitten TTS which is a a [34:05] subdirectory of kitten ML what is this [34:08] what is this ML so this is text to [34:11] speech model uh which people can host u [34:15] self host it's only like 25 megabytes [34:18] smallest one this is uh I'm a big fan of [34:22] open source models. Um and in term uh [34:25] that's great for privacy privacy fox [34:28] people and in general this model what [34:32] gives you is you know sometimes we go [34:34] articles and there is listen to this [34:36] article and you can do this with this [34:40] model and without paying anyone you can [34:43] just host in your WordPress website in [34:45] your hosting hopefully you are not using [34:47] share hosting um so and so this will be [34:51] enough resources to run this but still [34:54] um perhaps it's impossible to sh run it [34:57] on share hosting and 25 megabytes quite [35:00] small and does a really good job um [35:03] converting your articles into tech and a [35:06] speech and comes with I believe eight uh [35:10] voices. Yeah, I just want to I I think [35:13] this team done a really great great job [35:16] getting this out and supporting open [35:19] source models and yeah so it says on the [35:24] the GitHub repo kitten TTS is an open [35:26] source realistic texttospech model with [35:28] just 15 million parameters. That sounds [35:31] like a lot to me but presumably it's [35:33] not. Um designed for lightweight [35:35] deployment and high quality voice [35:37] synthesis. I confess I haven't used [35:39] anything like this largely because every [35:40] single thing that I do is actually me [35:42] speaking. But I do see this more and [35:45] more on websites wherever I tend to [35:47] roam. There's quite a lot of here's the [35:48] text and then here's the accompanying [35:51] usually AI generated I don't know audio [35:53] file which is almost like a just almost [35:56] like somebody reading it out. I um I [35:59] find them quite hard to listen to most [36:01] of the time because they don't sound [36:02] real, but maybe this is [36:04] >> there's there's some examples down [36:06] there. If you scroll, you will see [36:08] there's a they put some examples where [36:10] they read the text. [36:13] >> Yeah, there's all different kind of [36:14] voices you can choose and experiment. I [36:16] think what's great about it like you [36:18] don't have to rely on third parties and [36:21] and also I'm uh in terms of [36:24] accessibility cases like you people can [36:26] just listen to article [36:29] and yeah okay then I will I will link to [36:33] it in the show notes if you have a need [36:35] to put you know speech on your blog [36:37] posts and things like that inside your [36:39] WordPress blog this will do that for you [36:41] by the looks of things. Uh, I'll just [36:44] Yeah, the the URL is not that difficult [36:45] to read, so I will just read it out. I [36:47] don't normally from GitHub because it's [36:48] hard usually. So, it's [36:49] github.com/kittenml/kitten [36:53] tts. So, that was the first one. [36:56] >> I was so disappointed it wasn't what I [36:58] thought it was. I have never clicked a [37:00] link on quicker expecting it to be [37:03] realistic meows. [37:05] >> Oh, I see. Okay. [37:06] >> I was I was I was all in for this. There [37:09] was going to be this realistic set of [37:11] meows coming out of it and I could [37:12] program my me some me some cat purring. [37:16] >> No, [37:17] >> there's a whole there's a whole [37:19] subculture of people. [37:20] >> I want a large language model that has [37:22] been trained on meows and purr [37:25] >> translated into cat. Yeah. Okay, that's [37:30] okay. So, there's the first one. Uh [37:32] second one from uh Onar is well I [37:36] presume this is your own because it's [37:37] the spam. You've got an annual report as [37:40] well. You've got some key findings here [37:42] but just just sort of run us through the [37:44] the highlight items of [37:46] >> Yeah. Yeah. I'm not going to take spend [37:48] too long on you on this. I just uh this [37:52] is usually we do every year where we [37:55] take some sample of data and look at the [37:58] data and this is a very actually related [38:01] to what Cloudflare um posted recently [38:04] about their um denial of service attack. [38:09] uh in recently we've seen a lot of um [38:12] attacks are coming from residential IPs [38:14] meaning like people they don't know [38:17] their devices are compromised like they [38:19] have smart boxes at home or they have [38:22] like their computers they their smart [38:25] devices they all like spreading u [38:29] starting spam campaigns or doing card [38:31] testing attack people are income are [38:34] very aware of this issue card testing [38:36] attack where people steal other people's [38:38] those cards, credit cards and they test [38:41] it on your WooCommerce website [38:43] and yeah this is about report about that [38:47] and yeah in general um awareness about [38:50] residential proxies [38:52] and yeah we seen again this year we saw [38:55] more attacks in WooCommerce and and also [39:00] WordPress is more popular so it has it [39:02] has seen more spam campaigns targeting [39:05] WordPress [39:07] um Yeah. Um that's all I Yeah. I [39:11] encourage people to read about it and [39:13] there is some um also free ways free [39:16] tools how they can protect themsel from [39:19] the attacks and against [39:23] >> yeah I will link to that again in the uh [39:26] in the show notes. Let me just see the [39:27] UR the URL is easy enough to say so I'll [39:29] read that in. [39:30] >> I think quite a lot of people who [39:31] consume this podcast they do not um they [39:34] don't read the you know the associated [39:36] posts. So, it's quite good if I can read [39:38] the URL. So, it's uh oops spam. So, [39:41] oopspam.com2025-spam- [39:47] report. So, you can go and check that [39:49] out and there are some key findings [39:50] there. We'll get on to a piece a bit [39:52] later if there's time from Cloudflare [39:54] and one of the things that they [39:55] particularly mentioned was the um the [39:57] use of uh smart TV [40:00] um hardware as one of the things which [40:02] has sort of ramped up I think over the [40:05] last quarter. We have a like a little [40:07] thing which we plug in the back of the [40:08] TV. It's I think it's Google TV [40:11] something like that. [40:11] >> Yeah. Also fire. [40:13] >> Yeah. [40:16] >> Okay. Right. In which case I think [40:18] you've got a third one as well. This [40:20] one. Um so this one again from Onar. [40:24] This one's called Well, it's on the MI I [40:26] don't know if it's MIT Sloan or MITS. I [40:29] don't really know. I think it's part of [40:30] the MIT uh infrastructure. It's the [40:32] Sloan school which has something to do [40:34] with I guess management or something [40:36] like that. Uh the title is open and [40:38] sorry AI open models have benefits so [40:41] why aren't they used more widely? Is the [40:44] answer to that because they're not as [40:45] good? Is that basically what it is at [40:48] the moment? Would that sort of [40:50] >> article also mentions about that um [40:52] there's usually 90% [40:56] quality difference. So usually and even [40:59] like uh even open source models they [41:01] close that gap very like in within 13 [41:04] weeks [41:06] >> and uh these models in the beginning [41:08] they were only like u open AI and all [41:11] the closed sources models so people more [41:15] got used to those models and people [41:17] started integrating to into their system [41:20] but then later there's more smarter [41:22] there's more smarter and equally yeah [41:26] high quality models also came out open [41:28] source they are yeah I think now it's [41:32] really expensive for companies to switch [41:34] I think that's one reason um people [41:37] don't want to deal with it but there is [41:41] and some people are concerned about open [41:44] source models because of privacy but [41:46] then there is open source models you can [41:48] host it in your own infrastructure if [41:50] you're a company and that's uh will be [41:52] told privacy and you own all your data [41:55] >> yeah and I think another concern they [41:58] raised there is um top open source uh [42:02] models now currently all made in China [42:06] >> and so they all come from yeah and [42:09] that's another privacy concern but again [42:11] you can host this uh this all the open [42:15] models locally and without any internet [42:18] connection [42:20] so that will solves the privacy issue. [42:23] Yeah. Uh in general the article goes [42:25] through why people should switch and [42:27] they should consider it and of course [42:29] there's a huge cost and yeah I um I have [42:34] this feeling that the year 2026 I'm just [42:37] taking this data from people out in the [42:39] wider ecosystem you know tech journalism [42:41] more broadly so not really WordPress [42:43] related and many of them seem to be [42:45] talking about the the real financial [42:48] upside down pyramid basically this sort [42:50] of really difficult to maintain [42:53] financial cial structure that a lot of [42:54] these AI companies are built upon. So, [42:57] you know, venture capital money, which [42:58] at some point the venture capitals will [42:59] wish to have back and and how the a the [43:02] AIs that many people have become [43:04] dependent upon will become significantly [43:06] more expensive for the exact same thing. [43:09] I do wonder as well g given the kind of [43:12] stuff that I would ever use AI for, [43:14] which might be literally just [43:15] summarizing a URL, you know, poke a URL [43:18] at it and say summarize that in a [43:20] hundred words. I I'd be I'm sure that [43:22] some of the open source ones would be [43:24] completely capable for me at that kind [43:27] of thing. Obviously, if you're trying to [43:28] do the cutting edge um coding and things [43:32] like that, then I'm sure that you do [43:33] need to reach into your pockets, but I [43:35] don't know, maybe open source will win [43:37] this in the end, but let's wait and see [43:38] if they're capable of using accessing [43:42] internet resources and using different [43:44] tools. Yeah, they much Yeah, they're [43:46] very capable nowadays. it just they [43:49] there's not as as much commercial [43:51] interest in open source models. So [43:53] there's not as much PR team behind them. [43:55] So I know [43:56] >> it's kind of kind of interesting how a [43:58] lot of the people in the open source [44:00] arena [44:01] um who advocate massively for open [44:04] source are quite happy to use these [44:06] proprietary technologies to uh you know [44:09] to build this that and the other thing [44:10] on top of WordPress. It's kind of [44:12] interesting how that whole [44:13] >> Yeah. You don't even have to like you [44:15] don't even have to host them yourself. [44:18] You can for example the open router um I [44:21] don't know you heard about this open [44:22] router is a infrastructure that hosts [44:24] all the different open-source [44:26] um products uh open source models and [44:30] you can use them through the companies [44:33] too and you get privacy benefits and [44:35] it's much much cheaper to use them to [44:38] for especially small uh simpler tasks. [44:42] So Ross, who uh who is in the comments, [44:45] sorry uh Tim, I'll just read Ross' [44:47] comment quickly. Um it surprises me says [44:48] Ross. Um the amount of the amount that [44:51] the WordPress project is talking about [44:52] AI so much but not talking about open [44:55] models. Sorry Tim, I think you were [44:56] trying to say [44:57] >> I was just going to put in a couple of [44:58] points. One is that there are two [45:01] different costs. open open source models [45:04] you have the cost to train the models [45:06] and the size that and infrastructure you [45:09] need to train the models and that is a [45:12] hardware problem that we just there [45:14] aren't companies going hello I'd like to [45:16] I have I have a data center over here [45:18] that I wish to train the models on I [45:20] will then hand that that open model back [45:22] into the real world and give it away for [45:25] free with the exception of unless that [45:28] is literally a company strategy or a n a [45:32] nation state strategy which is the [45:34] Chinese model process which is we're [45:36] going to throw this out as a disruptor [45:39] and that's [45:40] >> oh [45:41] >> really useful for us you know people who [45:43] are using the open source model they're [45:44] not doing it because it's an for open [45:47] source in the nice way we think of [45:50] they're doing it to disrupt against the [45:52] their larger competitors um and to stop [45:55] necessarily proprietary models taking [45:58] single hold if you do not have someone [46:00] who has enough hacking to give you the [46:03] data center to train these models. It's [46:05] really difficult. Then we have using the [46:08] models themselves. They still require on [46:12] the whole significant hardware. You were [46:14] saying, "Oh, can I summarize something?" [46:15] Absolutely. [46:16] >> Yeah. [46:17] >> Um the uh I have a a Quen model on my [46:21] laptop. Absolutely perfectly fine for [46:23] summarizing. takes about if I f if I [46:27] pointed at a uh an article it takes [46:30] about 45 to 50 seconds to come back to [46:32] me with a summarization [46:34] >> right right which is [46:35] >> I if I pass that to chat pt it come back [46:38] in 5 seconds [46:39] >> yeah okay there's the difference right [46:40] that's the [46:41] >> quality would be there that's fine for [46:44] one article scaled once on my machine [46:48] the problem comes when you want to do [46:50] heavy lots of workload so let's say I'm [46:53] doing I want to audit some code [46:56] um with a open- source models as we have [46:59] now let's even if it's a relatively [47:02] small code base [47:04] with my current hardware which I don't [47:07] have I'm not a massive company with lots [47:09] of hardware but we just hardware that I [47:11] have currently I can do it it just takes [47:14] a really really long time [47:16] >> whereas if I threw that same code base [47:19] at claude code it would be done in a [47:22] fraction of the time at that extreme [47:24] cost. I can throw my code base via [47:27] something like open router, but actually [47:30] open router is quite expensive because I [47:33] have to make the if I'm making a per API [47:35] request. I'm I'm still paying for their [47:39] hardware at much more inflated rate [47:42] because they have to make money. Open [47:44] routters is not a charity. They make [47:46] money on those requests through. So we [47:48] haven't yet reached a point where we [47:50] have models that are small enough to run [47:52] locally at performance that's required [47:54] to make the open source a viable model [47:57] and what we don't have yet is model a [48:00] way of training models that's small [48:02] enough that they or we can distribute [48:05] well enough um we we we will I'm sure [48:10] see in the next sort of 6 months to 12 [48:12] months um seti style setups if you [48:16] remember the oldi Yeah. Yeah. Hive mind [48:19] kind of thing. [48:20] >> Yeah. I'm sure that we will that's the [48:22] route that open source will ultimately [48:24] go down to solve the distribution [48:27] problem at making that. But to do that [48:31] introduces [48:34] amount of complexity, [48:36] security implications and pollution of [48:39] training data. Also just the whole how [48:42] do we train these models on what do we [48:44] train these models? Um because one of [48:46] the things about the open source models [48:48] coming from China is that there's at [48:49] least allegations that they're trained [48:51] on other AIs um which and the quite [48:55] often the training methods for these [48:56] larger larger open source models isn't [48:59] as transparent as we'd like. We get the [49:01] end result as an open source model but [49:03] how it got there is sort of like fuzzy [49:06] on the way through. Um that is the title [49:10] of today's podcast but there's no way [49:12] that I can sound like that. I want to [49:14] say the word fuzzy. [49:20] >> Make it make it in a really small font. [49:22] >> Tiny font. Yeah, tiny. [49:24] >> Tiny font. [49:25] >> Um, okay. So, that's really interesting. [49:27] Ross has got just sort of to we put that [49:30] comment up, but I'll just put it up one [49:32] more time. Um, that was the context for [49:35] Tim's reply. That is so interesting. If [49:37] you've got this intuition that some sort [49:39] of federated version of AI like uh like [49:44] the SETI thing that you mentioned where [49:45] people are sort of I don't know letting [49:48] their CPUs in idle time do some grunt [49:51] work in the background and and [49:53] contributing to a [49:54] >> probably better than turning it into a [49:56] crypto miner which is what those CPUs [49:59] are currently being used whether they [50:00] wanted to or not. [50:02] >> Yeah, I couldn't agree more. Um, I [50:04] however don't have access to a data [50:06] center, but just you know because it's [50:08] good news. I do have access to loads of [50:10] grapes and um and I'm going to eat these [50:14] grapes. My ambition during the course of [50:15] this podcast is to finish all the grapes [50:18] and uh and then really need a week. Just [50:21] to tie it back into security for a [50:23] second that the uh one thing we are [50:24] seeing more of is that um bad actors [50:27] when they are compromising servers are [50:29] doing so to get the CPU usage for AI [50:34] stuff for writing out spam emails and [50:36] blahy blah. Um not to the same it's only [50:41] in small pockets at the moment but I'm [50:42] sure we're going to start seeing. So if [50:44] you want to look for where we're seeing [50:47] uh distributed uh AI at the moment, it [50:51] is actually being used more as on the [50:54] aggression side. It is being used by bad [50:55] actors, not necessarily [50:58] for the benefit of mankind in general. [51:00] >> Who couldn't have seen that coming? I [51:02] know you're shocked. [51:03] >> Yeah, what a shame. I'm just going to [51:05] move I'm going to flick to something [51:07] that I was going to do at the end [51:08] because you were talking about um the [51:10] web being scraped and the way that you [51:12] know the open source models might scrape [51:14] it just as an aside. I was going to [51:17] leave this till the end. This is over on [51:18] the tech dirt blog. It's called [51:21] preserving the web is not the problem. [51:23] Losing it is. And this is just a curious [51:26] byproduct of that scraping that um [51:29] institution the institution which we've [51:32] all come to know and love the way back [51:34] machine. So the internet archives way [51:35] back machine is really caught at the [51:39] minute as to how best to deal with AI [51:42] agents scraping it because um it's being [51:46] used largely by AI agents scraping it [51:50] and a lot of the content for example [51:51] from institutions like it mentions a few [51:54] the Guardian the New York Times and [51:57] Reddit [51:58] um are now blocking the access to the [52:01] internet archive to scrape future [52:03] content. content, maybe past content as [52:05] well. I don't really know how that would [52:06] work, but future content. Um, because [52:10] there it's a way for the AI agents to [52:13] sort of get round their content [52:14] restrictions and what have you. And so [52:17] that would be a shame. That would be a [52:18] real shame. I kind of I didn't foresee [52:20] that one. You know, like I said a minute [52:22] ago, you could foresee the AI um, you [52:25] know, the miscreants misusing AI to [52:28] write spam emails, but for some reason, [52:30] I didn't see that one coming. And that [52:32] just seems like a a real shame if [52:34] something like that was to happen. Mike [52:36] Maznik, I think, wrote this. No, it [52:38] wasn't actually. It was Mark Graham. [52:39] It's not very long. It's pretty quick to [52:40] read. So, go and check it out. But it is [52:42] a bit of a shame. [52:44] >> The Internet Archive has for has its own [52:46] set of controversies based on scraping [52:50] when it may or may not be allowed to. [52:51] So, there is something ironic about [52:53] this. [52:53] >> Yeah. Yeah. Yeah. That's right. It's uh [52:55] shouting fire in the theater or [52:57] something like that. Yeah. Yeah. Yeah. [52:58] Okay. Um Okay, let's move on. Let's do [53:00] some sort of WordPressy stuff. So, this [53:03] is the first one. So, WordPress 7.0. [53:06] Beta 1 is just around the corner. I [53:09] can't actually remember the day it [53:10] drops. Maybe it's dropped already. Can't [53:12] remember. Um, but anyway, it's coming. [53:15] Um, and so it's time we go through this [53:17] sort cycle of beta releases and then [53:20] finally the the actual release will come [53:22] fairly soon. Time to start playing. [53:24] Don't put it on a production site, but [53:26] there is some fa there is some pretty [53:28] good stuff in this release. If you want [53:30] to follow somebody who's been following [53:32] the collaborative editing side of [53:34] things, go follow Steve Burge from [53:37] Publish Press because he more or less [53:39] daily uh on X tweets about his latest [53:43] discovery um to do with this. But I'm [53:46] taking my guidance from him, but it [53:49] seems like the whole collaborative [53:50] editing, which many people thought was [53:52] going to be a real roadblock, you know, [53:54] a really difficult thing to pull off, it [53:55] feels like it's it's coming. Um and it's [53:58] going to drop. We'll see what kind of [54:00] hardware you need to make that work. Um, [54:02] and then there's another whole there's a [54:04] whole bunch of stuff around design [54:05] improvements. So, the video uh embed [54:08] block and the grid block and the icons [54:10] and the breadcrumbs and the headings [54:11] blocks have all had adaptations as have [54:14] the nav as has the navigation block. And [54:16] then there's a bunch of developery stuff [54:18] as well. You can read it all in this [54:20] piece. Um, it's called WordPress 7.0 [54:23] beta 1 and it's on the WordPress.org [54:25] news blog. So, go check that out and go [54:29] and have a play a bit like Steve does. [54:31] You three, anything about that before I [54:33] move on. [54:34] >> Don't run it on production. [54:36] >> Don't run it on production. Look, it [54:38] literally says it on the first line. [54:40] >> You know, unless you're me, in which [54:42] case you've updated your site to use it. [54:44] But for the rest of you, don't run it on [54:46] production. [54:46] >> Yeah, that's right. Just don't. Um, so [54:49] there we go. Oh, and look, there's [54:51] Steve. Ha, there he is. Uh, hi, Steve. [54:55] Um, I was just saying that everybody [54:57] should follow you if they want to learn [54:58] about the collaborative editing because [55:00] you every few days it seems you drop a [55:02] little nugget of something that you've [55:04] played about within WordPress 7 and uh, [55:06] yeah, I'm very grateful for you doing [55:08] that. Appreciate it. So anyway, go and [55:10] check that out. Just a quick mention, [55:13] uh, friend of this podcast, he's been on [55:15] it lots and lots of times, no doubt will [55:17] be in the future as well, Remus Dice. [55:19] Um, he's launching a new service. He's [55:22] he's had a service for a little while. [55:24] uh truer than north. But I I won't try [55:27] to pause this. It's a fairly lengthy [55:29] written article, but he is um sort of [55:31] pivoting what he does in the WordPress [55:33] space um and picking up the mantle that [55:36] Guten sorry Gildenberg put down the [55:39] project run by Jonathan Wald and Tammy [55:41] Listister um which they put on ice I [55:44] think just before Christmas something [55:46] like that. Um I don't think Remus is [55:49] literally directly picking it up. But as [55:51] it's gone out of existence, he can see [55:53] that there's a sort of um a hole in the [55:56] market there. So he's going to be trying [55:57] doing what they're doing. In other [55:58] words, communicating with people, trying [56:00] to figure out where they fit with their [56:02] products and services in the WordPress [56:04] ecosystem. So he's going to be like a [56:06] consultancy for that. He does however [56:09] have a direct rival because I spoke to [56:12] Matt Cromwell uh during the course of [56:14] this week. Um and he co-founder of [56:17] GiveWP, worked for Stellar WP for a long [56:20] time. So equally credible in terms of [56:22] their length and breadth of history in [56:24] the WordPress space. Um he's launched [56:27] something called Roots and Fruit which [56:29] looks like it's going to be exactly [56:32] similar very analogous to what Remus is [56:35] doing. So, if you need a bit of guidance [56:37] in your WordPress company's journey, [56:40] check out Rimkus' blog, but also if you [56:42] check out um the Tavern podcast that I [56:45] did with Matt Cromwell or go to [56:46] rootsandfruit.com, [56:48] you'll be able to see what they're both [56:50] up to. Anything about that? [56:51] >> In the pre-show, we were suggesting that [56:53] they should we should just set them up [56:55] in a ring and fight, but I've [56:56] subsequently thought that given that how [56:57] unfair that would be to Matt [56:59] >> Yeah. [56:59] >> that a rap battle might be in order [57:01] instead. [57:02] >> Oh. um rap battle. [57:05] >> Yeah. Could Could we not do both? Would [57:08] it be possible to have them both doing [57:09] that at simultaneously? A a sort of cage [57:13] fight meets rap battle all at the same [57:16] all at the same time. Uh I apologize [57:19] Matt. This wasn't my idea or Remus. It [57:21] was Tim's idea. I'm just saying. [57:25] >> But but Nathan, you did fully support it [57:27] before the show. So [57:28] >> No, no, no. I I don't remember any of [57:30] that. I can't I can't can't remember [57:32] that at all. No, no, no. Yes, it would [57:35] be fun. Um, and then Taco's here. So, I [57:38] know Taco knows Remus very well. Maybe [57:40] you could persuade him to get in a [57:43] cage fight with Mar. This is not [57:46] >> I don't think anyone I don't think [57:48] anyone wants to get in a cage fight with [57:51] Remus. I don't know. I [57:54] >> no there's there's probably like a [57:56] handful of people on earth who' be who [57:57] best roads I think there's not many and [58:01] they probably all live in like Norway or [58:03] something like that. [58:04] >> I'm going to go back to the rap battle. [58:07] >> Yeah. Okay. All right. So, uh Remus and [58:10] Matt, if you could uh just crack on with [58:12] that, organize the rap battle and then [58:14] we can report back on it. I [58:15] >> I feel like a major word camp that would [58:17] work nicely. [58:18] >> Oh, yeah. Yeah. see who's like the best [58:21] at advising clients. Look, I tried to do [58:24] some sort of rap gesture then. I sort of [58:26] did this. Don't know. Don't know. [58:29] >> I I feel like that would be the [58:31] cringiest thing ever. [58:33] >> Yeah, that's why it should be done, [58:35] Mark. That's exactly. [58:36] >> Absolutely. Absolutely. Be it would be [58:38] the most watched WordPress TV thing [58:41] ever, probably. [58:42] >> Jamie Mars can compare it. It could be [58:45] Oh, yeah. Oh, yeah. I think I think I [58:47] think we've cracked it. So yeah, Taco's [58:50] saying poor Matt. A not poor Remus, [58:53] though. I know she's saying, but Amber [58:56] likes it. She's got a giggle icon. And [58:58] frankly, if Amber likes it with a giggle [59:00] icon, it has to happen. That's all there [59:02] is to it. Keep it coming. This is This [59:04] is brilliant. We've lost Onar. I don't [59:06] know if we've lost Onar or Onar's camera [59:09] has sort of just died the death, but [59:11] we've we've lost Well, no, we haven't [59:13] lost him. He must still be there in some [59:14] capacity because I can see that he's [59:16] still on the screen. So Onar, if you can [59:18] hear me. [59:19] >> Sorry, I'm here. I just have a problem. [59:22] >> That's fine. If the camera's busticated, [59:24] no problem. Just stick with us for the [59:25] audio. [59:26] >> I'm listening. I'm here. [59:27] >> Yeah, that's great. Thank you. Okie [59:29] dokie. Let's move on a little bit and [59:32] have a look at this. Last week we [59:34] reported on the fact that um the WPCC, [59:38] now annoyingly there are two acronyms [59:41] that fit WPCC. [59:44] There's the WP [59:46] what is it? community credits or [59:48] something like that. And then there's [59:49] the WP community collective. I can't [59:51] remember. [59:51] >> I thought you were talking about [59:53] Wimbledon Wimble Wimbledon's public [59:55] cricket club [59:56] >> that Well, there's three now apparently. [59:59] Uh but that one doesn't often feature on [60:02] this show sadly. Uh but it should. [60:05] >> Sorry, it was it's Wimbledon and Putney [60:07] Commons was the is it along with the [60:10] parent council. Honestly, it's been too [60:13] long since we talked about them on this [60:15] podcast, but here we are. So, WPCC in [60:19] this case means the WP WP Community [60:23] Collective. Uh, last week we reported [60:25] that um one of the founder members and [60:27] board members um had stood down. Uh that [60:31] was Katie Adams Farrell as you can see [60:34] if you can see the screen. Um quick off [60:36] the news, the the position has been sort [60:39] of filled if you like. There's uh a [60:41] couple of people that have stepped up, [60:43] but Chris Reynolds is probably the most [60:45] notable one to mention. He has stepped [60:48] in as the interim uh president. The [60:52] other one is uh Jono Alderson [60:55] and uh yeah, so they're both going to be [60:57] acting as part of the WPCC uh endeavor. [61:00] I know Courtney's joining us in the [61:02] comments. She probably has a lot more uh [61:04] sense to say about this because she is [61:06] one of the founders as well. Um and [61:10] yeah, the endeavor of this project is to [61:12] is to you know sort of grow the [61:14] community in a fair, equitable, [61:17] transparent way. And uh I feel I've had [61:21] a few chats with uh Chris Reynolds. I [61:23] don't know that I've ever spoken to [61:24] Jonno Alderson, but my the only [61:27] experience that I've had with Jonno was [61:29] watching him speak um at a WordPress [61:32] event. And boy, he is a force of nature, [61:34] isn't he? He's really quite quite [61:36] something. he was able to speak on the [61:38] subject of SEO and it was really it's [61:40] powerful watching him. So we'll have to [61:42] see we have to wait and see. Tim's sort [61:43] of having a bit of a bit of a giggle [61:45] there. [61:45] >> I just like the idea that yes, Jonno is [61:48] a force of nature and we should leave it [61:49] there. [61:50] >> Yeah. Okay. Okay. [61:51] >> Just because out of context that sounds [61:53] amazing. I think Jonno would appreciate [61:54] that as a statement as well. [61:56] >> Yeah. So Chris has got a sort of fairly [61:59] muted like fairly political way of of [62:03] explaining what he wants to do as part [62:05] of his presidency. But um Jono is a [62:08] little bit more forthright. He says [62:12] WordPress so this gives you some idea of [62:14] maybe where Jonno wants to take it. Uh [62:17] WordPress runs on contribution but too [62:18] often the people and organizations doing [62:20] the work are expected to absorb the [62:22] cost. Alderson said, quoting him again, [62:24] "The WPCC is creating real structure [62:27] around sustainability, accountability, [62:29] and contributor well-being. I'm proud to [62:32] be joining the board and helping to push [62:34] WordPress towards a healthier, more [62:36] equitable future." This is on the [62:38] repository uh [62:41] blog, so you can go and check it out [62:43] over there. Uh anything from you three, [62:46] excuse me, about that? [62:47] >> Not about that, but can we just take a [62:49] moment? So, uh, you can't see this on [62:50] the podcast, but the amazing picture [62:52] that the repository found of Chris, [62:55] which is like you can't even see his [62:58] face. He's just [62:59] >> Which one is it? Is this him here? [63:01] >> Yeah. [63:02] >> Maybe he um maybe he assumes the camera. [63:05] He sort of lurks elsewhere. It's not [63:07] possible to find pictures of him because [63:10] whenever the camera comes out, he [63:11] disappears into a corner. [63:14] It looks like um like one of those [63:16] pictures that would be taken, you know, [63:17] you'd see on the front page of a of a [63:19] newspaper. I feel like that's right it [63:21] really well with the black and white and [63:23] it made me smile. It's like, yep, that's [63:25] that's the person. [63:27] >> Um let's see what Reese says. John Jonno [63:31] is a little bit more forthright is is [63:33] the understatement of the century. Okay. [63:36] Um I meant in that article in particular [63:38] because Chris said something and then [63:40] Jonno said something and John [63:41] >> Oh, no. No. life. Jono. [63:45] >> Yeah. Yeah. Uh, Taco says, "It's like [63:49] Jonno Fest." Jonno is absolutely bloody [63:52] brilliant, says Taco. Okay, [63:55] that's it. Ideal. Uh, D. And then I've [64:00] got some commentary here about [64:01] subscribing to my newsletter. [64:04] Uh, I don't know. I'll send me an email [64:06] and I'll figure it out. It's [64:07] adminwpbuilds.com. [64:09] He [64:10] >> was talking about the tavern, not you. [64:12] Oh, is he tavern? [64:13] >> Yeah, [64:14] >> I have no control over that, but send me [64:16] an email anyway at that email address [64:18] and I'll get it looked at by the team [64:20] that run the uh the Tavern website. [64:22] Okay, thank you very very much. Anyway, [64:24] there we go. Uh Courtney saying, [64:26] "Welcome to both Chris and Jono joining [64:28] the WPCC." [64:31] Uh oh, maybe it's one of Zach's. Zack [64:34] Zack, if you've ever met Zack, Zach's [64:36] like a total pro photographer. He takes [64:38] photographs of bands, mainly metal bands [64:42] for I think it's a side gig. It's not [64:44] like exactly what he does. He obviously [64:46] works in the tech space. Um, but side [64:49] gig. So maybe that is that one of yours. [64:51] Uh, Zach, did you actually take that [64:52] picture? That's interesting. Right. [64:56] Okay. There we go. That's that piece. [64:57] Right. Let's move on. I'm not [65:00] Unfortunately, I never gave this enough [65:02] time to really pause what was going on [65:03] here. I read it ever so briefly, thought [65:05] I'd understood it, and now that it's in [65:07] front of me, I'm not sure that I did. [65:08] So, I'll just I'll just say what I think [65:11] is going on here, and then I'll let [65:13] either the hive mind of the comments or [65:15] you three panelists tell me that I've [65:17] got it wrong. So, Matt Mullenweg took to [65:19] the make.wordpress.org Slack this week. [65:23] Is that right? Yeah. 18th of February. [65:26] Uh and said, "I've heard that w so word [65:28] camp Asia, Europe, and uh US have [65:32] separate Slack groups." Um but it would [65:36] be better to have everybody in the [65:37] main.org group. And then he says, "Let's [65:40] move everything over." He says, "I think [65:42] this was done because there was too much [65:44] friction in making new channels. Totally [65:45] fine to create public channels for big [65:47] events and private ones for planning, [65:49] etc. let's deputize a dozen people to be [65:52] able to spin up channels or admin them. [65:55] And so I don't know if this is a thing [65:57] which is now done or happening or what [66:01] the need was, but there was a there was [66:03] a few comments that came in saying it's [66:06] quite handy to have those things [66:08] separated in it. For example, at Word [66:10] Camp Europe, they like to keep it secret [66:12] where the um where the next Word Camp [66:14] Europe is. That's part of the whole [66:15] thing. you know, you show up and then [66:17] the final day you get to find out with [66:18] the with the expensively made video [66:21] where it's going next time. And [66:22] obviously, if everything was in a public [66:24] channel, that might not be the case. So, [66:26] I'm not entirely sure what the um what [66:29] the reasoning behind this is or whether [66:31] it'll happen. But the other thing which [66:33] is kind of interesting here, certainly [66:34] from my point of view, it says here, I'm [66:37] very open to water cooler or more [66:40] socials channels on Slack as well. We [66:42] should have some fun there as we work [66:44] together. So it's not just his endeavor [66:47] here is not just to have WordPress, you [66:49] know, code and events, but have other [66:51] silly things like a dog's channel. I [66:54] mean, it literally this is what's on the [66:56] page. A dog channel, a cat channel, a [66:59] games channel, a self-promotion channel, [67:01] and a Today I Learned channel. Now, [67:04] obviously, if you listen to this [67:05] podcast, you know I just like messing [67:07] about. So that that's music to my ears. [67:11] Let's have as many Let's have a hamster [67:13] channel. Let's begin there, shall we? [67:15] So, anyway, I don't know if this is big [67:17] news to you. If it's of no interest [67:19] whatsoever, I'll just leave it. If you [67:21] want to talk about it, we will. If not, [67:22] I'll just move on [67:25] whilst I'm eating grapes. [67:28] >> That's a big fat no. Nobody wants to [67:30] talk about that. Okay. Did any comments? [67:33] >> The only thing I have to say is Yeah, I [67:35] agree. We need to slack better. I'm not [67:37] going to say anything beyond that. So [67:38] the the headline is correct because [67:42] there's there's a lot of clutter in [67:45] Slack and and all the [67:48] social uh chat uh platforms. So I I [67:53] would like for there to be something [67:54] somewhat centralized, but I'm not sure [67:57] how it should be done. [67:58] >> Yeah. Yeah. But but just kind of [68:00] interesting. I didn't I suppose it it I [68:04] just don't know what the thought process [68:05] is behind it in terms of whether it's [68:06] needed or not. Um, but it's been mooted. [68:10] You can go and join in the comments that [68:12] there are quite, [68:13] >> you know, I mean there's not lots and [68:15] lots and lots of comments, but there's [68:16] quite a few comments. [68:18] >> I will be curious to know what they talk [68:19] about in cats and dogs. [68:21] >> Yeah. Yeah. [68:22] >> It's just cat videos. Cat dog videos. [68:27] >> Yeah. The cat cat channels got to be [68:29] more than dogs. [68:32] >> The dog channels exclusively cat for [68:34] cats. By the way, um, somebody's ears [68:37] were getting hot. You know, somebody was [68:40] could hear us over the internet. Remus [68:42] has joined in. Um, [68:45] >> yeah, I agree, Remus. I think that Matt [68:47] would hands down beat you in the [68:49] >> I was going to say, what's the [68:51] underlying What's What are you trying to [68:52] say there, Remus? What? I think you're [68:55] implying something, aren't you? You [68:57] know. Okay. What if What? What if? I'm [69:01] not even going to say it. It's so [69:05] what if Matt was allowed a small [69:07] rounders bat or something like that to [69:10] sort of offset the [69:12] >> You keep going to the violence rap [69:15] battle. That was the suggestion. [69:17] >> Never done violence before. [69:21] >> It's so funny though. I'm absolutely [69:23] loving I'm loving the rim because it's [69:25] even bothered to comment though. That's [69:27] absolutely brilliant. Um so anyway, [69:29] we'll see see if that happens. [69:32] I'm surprised he didn't put the word [69:33] fight in quotes while he was at it. [69:36] >> Yeah, that's right. Okay, let's see if [69:39] Matt Cromwell drops in and and offers it [69:42] offers it off. Okay, right. Let's get [69:44] into some sort of security stuff. [69:46] There's a few bits and pieces here. Um, [69:48] it's totally out of my pay grade this, [69:50] but I did think this was kind of [69:51] interesting. So, this is talking about [69:53] AI and security all in one go. So, [69:56] somebody that I followed on social media [69:58] for a little while, this is on [69:59] Masttodon, uh Kevin Bowmont, he says, [70:02] "Today in infosc job security news. I [70:06] was looking into an obvious [70:07] vulnerability introduced into a major [70:10] reb web framework today and it was [70:12] committed by username Claude on GitHub [70:16] vibe coded basically." So, I started [70:18] looking through Claude commits on [70:19] GitHub. There's over two million of them [70:22] and it's about 5% of o all opensource [70:26] code this month. As I looked through the [70:28] code, I saw the same class of [70:30] vulnerabilities being introduced over [70:32] and over again. Several a minute. [70:35] Honestly, I don't I don't think we need [70:37] to talk about that, but I just thought [70:38] that was an interesting thing to raise. [70:40] You know, just like what what what is [70:42] the future we're painting for ourselves [70:44] if if stuff like that is true? [70:46] >> Breaking news. AI was trained on bad [70:49] data. Bad input and bad data. [70:52] >> Oh, increases vulnerabilities. I'm [70:54] shocked. [70:54] >> No kidding. No kidding. Should we go to [70:56] the other interesting one that I found? [70:58] Where was that? Was it this one? No. Uh, [71:01] let me try. We'll come to that at the [71:03] end. Where's it gone? Where's it gone? [71:06] Where's it gone? [71:07] >> I think that one you were just on the [71:08] open the open AI has a open source has [71:12] an AI slot problem. [71:13] >> That was not the one I was thinking of, [71:15] but that is kind of interesting. No, [71:17] I've forgotten what the context of it [71:19] was. Maybe it was that one. No, I can't [71:22] remember. But there was one which sort [71:23] of jived perfectly with this one. Yeah. [71:26] Anyway, we'll come back to those in a [71:28] minute. But I just thought that was kind [71:29] of interesting. You know, we're painting [71:30] ourselves into a future where the robots [71:32] are now Oh, no. It was the one about the [71:34] robot. [71:35] >> Are you talking about the Sean the Sean [71:36] Hooper one? [71:37] >> Yeah, that's right. Which one was that? [71:39] Where? Let me see if I can find it. [71:41] >> It's uh when Rob When robots fight back. [71:45] >> Yeah, let me find it. You're right. [71:46] You've got the you've got the measure of [71:48] it. That's the one I'm [71:49] >> while you're looking for that. One of [71:51] the things that's interesting about that [71:52] is that these commits are being actually [71:54] submitted and are being submitted and [71:57] signed by Claude. So this is like hello [72:00] I we're not we're not even pretending [72:02] that there's someone looking over this [72:04] and have manually putting this in. This [72:06] is and this is just this is the AI has [72:09] submitted this. This is signed and [72:12] through. And there's part of me that's [72:14] really pleased that's h that's [72:15] happening. [72:17] because right now it's really difficult [72:19] if you get a piece of code in front of [72:21] you and you're like who who or what [72:25] coded this? Um so it would be nice to [72:29] actually see more AI code being uh [72:35] properly [72:37] um [72:39] offered by having showing who's actually [72:41] doing the authorship. Okay. [72:43] >> What the position of the human was [72:45] inside that. So actually I think we [72:47] absolutely know we are there not 5%. [72:51] It's going to be [72:53] >> masses more. It's clearly 5 5% of isn't [72:56] that's that's a that's not not the real [72:59] figure. It's going to be significantly [73:01] more. We in the space of just 12 months [73:04] the development space has completely [73:06] changed. If I'm not I reckon there's [73:08] significantly more than 5% has been of [73:11] just the WordPress co core code has been [73:14] committed. There's going to be way more [73:16] than that generated by AI in the last [73:18] commit set of commits for WordPress 7. [73:21] If you the real time collaborate if we [73:23] took one of those features the real-time [73:24] collaboration part and just looked at [73:27] how much of that was AI coded I suspect [73:30] that that is going to be quite high. [73:32] It's not going to be five. It's not [73:33] going to be 10. It's not even poss might [73:35] be even in 20 30s 40%. [73:38] But [73:40] you don't know that. [73:42] >> No, [73:42] >> because we have no easy way at the [73:44] moment of verifying other than looking [73:46] at it and going, "God, that's AI slop." [73:48] I'm not saying the real time stuff is AI [73:50] slop by the way that was probably [73:52] but that there is a but it gives you an [73:55] example. We at the moment we don't know. [73:57] So it would be good to have some sort of [73:59] ownership and some custody chain so we [74:02] can actually see what's generated that. [74:04] >> Okay. So you're saying that having some [74:06] huristics as to that portion was [74:09] definitely humans and that portion [74:11] wasn't having some reliable data on that [74:13] and people not lie. That's kind of [74:16] interesting. By the way, I never get to [74:18] say the word breaking on this podcast, [74:19] but I'm going to say breaking because [74:23] look, here it is. It's official. [74:27] Pat Cromwell has [74:30] >> he's joined in. [74:33] >> I I would also point out that Remicus' [74:35] reply of uh his his last reply [74:38] explaining that uh you know the dude [74:40] who's been doing Taekwondo for eight [74:42] years and kickboxing for free. And I'm [74:44] going to emphasize again in a rap [74:47] battle. [74:48] >> Okay. Okay. Where's the comment about [74:50] Taekwondo? Oh yeah. Okay. So just I mean [74:54] I'm not trying to [74:57] I'm really not but it just it's funny [74:59] right it's quite funny. Uh so Matt [75:01] Cromwell says my bets on the little guy [75:03] in that match and then Remus comes back [75:07] with mine are absolutely on the big dude [75:09] who's been doing Taekwondo for eight [75:11] years in kickboxing for three. Okay. I [75:14] think at this [75:15] >> I think the other thing that we have to [75:16] keep in mind here is Matt really is a [75:19] lover not a fighter. Okay. So there's [75:22] there's that, too. So [75:24] >> that's if if I um if I use that as the [75:26] title of this episode, it's going to be [75:28] misconstrued. Um if I use Matt's Matt's [75:32] really a lover, not a fight. I don't [75:33] think anybody's gonna pick off. [75:35] >> But it's true. Matt Matt, you know, Matt [75:38] would just go and give Remus a big hug [75:40] and that would be the end of it. [75:42] >> Okay, Taco, can I just say thank you? [75:44] Taco's now divorced us from all [75:47] responsibility. He's decided to take on [75:49] organizing it. That's great. We wash our [75:51] hands of the whole thing. It was a [75:53] terrible idea, but you Yeah, you can I [75:57] still submit a session? Yes, of course [75:59] you can. You do it. Let's uh let's get [76:01] them in a Let's get them in a ring. This [76:04] is hysterical. I love this. Uh and then [76:07] Matt, finally, rap battle. Okay, rap [76:09] battle it is to go with music majors and [76:11] native English speakers. Okay, I dare [76:14] you to use that title. [76:16] Lord, just [76:20] some episodes are easier than others. [76:22] Let's put it that way. I have to keep up [76:24] with the comments as well as trying to [76:26] make sense of everything that's coming [76:27] in. Okay, Elliot Richmond's joining us. [76:29] Hi, Elliot. Loving your blog, by the [76:32] way. It's very calming. Um, recommits [76:35] via AI. This is a very good reason for [76:37] AI governance ASAP. [76:40] How would that be done? Who would govern [76:43] AI? You'd have to have some sort of UN [76:45] for AI, wouldn't you? Because cross [76:47] >> EU EU has a bunch of laws around AI [76:51] safety and governance. [76:53] >> I don't know how that going. [76:55] >> I mean, the EU has a bunch of proposals [76:58] for laws around it, right? [77:01] >> But [77:01] >> I'm not sure they actually implemented [77:03] it or not, but [77:05] >> they they jumped on that train very [77:07] quickly. [77:08] >> We all know that governments are [77:09] excellent at do at understanding [77:11] technology usage. Oh yeah. [77:17] >> What team say about the you using AI [77:20] code comments? Uh I know that people [77:22] some companies I some I have some [77:24] friends in uh Shopify I think they also [77:27] publicly disclose that they use a lot of [77:30] AI comments but there's always human in [77:33] the loop that uh reviews all the pull [77:37] requests by um AI models and so you want [77:41] to make sure that uh all the codes are [77:44] align with your the standards your [77:47] internal standards and also there's no [77:50] vulnerabilities and stuff. [77:52] >> Yeah. I Yeah, I think if you just let [77:54] the AI to do pull requests and submit [77:57] the codes, then uh there will be a lot [77:59] of problems. [78:01] >> So, sorry, sorry. [78:03] >> Yeah, go ahead. Yeah, go ahead. [78:05] >> So, I didn't mean to interrupt. Sorry. [78:06] Um, but you said something which led to [78:09] the article which I was trying to find [78:11] which, uh, you got quite right, Mark. It [78:13] was the one and I've actually now Oh, [78:15] for goodness sake, I've now lost it. [78:16] There we go. It was the Shan Hooper one. [78:19] >> Uh, this came up this week. So Sean [78:21] Hooper um posted a piece called When the [78:24] Robots Fight Back. And I I'll just read [78:27] it because it's so short, but it's so [78:29] interesting. It says, "Today I saw [78:30] something I'd never seen before. A bot [78:32] taking revenge on a human. Matt's plot [78:35] lib is a popular open source Python code [78:37] library used to make visualizations. [78:39] Yesterday, a contributor named Krabby [78:42] Wrathburn submitted a bug fix. The pull [78:44] request was closed by Scott Shambbor [78:47] stating the project's rules that ban AI [78:50] submissions. Usually, this would be the [78:52] end of things, but no. This bot, which [78:55] runs the open claw AI, instead decided, [78:59] you can't make this up, decided to [79:02] create a blog post calling Scott out, [79:04] which it later retracted. [79:07] In my favorite part of this interaction, [79:09] the bot tells Scott, "Judge the code, [79:11] not the coder. Your prejudice is hurting [79:14] Matt plot lib. [79:16] What I mean what's happened to us? I [79:20] mean here we are talking about Matt [79:22] Cromwell and Remus having a fight. It's [79:24] this it's this it's completely hijacked [79:26] our minds. It's all it's all ahole. [79:30] I mean, seriously though, getting back [79:32] to that piece that isn't that just nuts? [79:35] Doesn't that like isn't every bit of [79:36] your senue going where how have we got [79:39] to this point where a a robot can can in [79:43] effect get the hump and then go off and [79:46] I mean obviously it's doing what humans [79:48] do, right? This is the sort of churish [79:50] petty behavior that humans exhibit and [79:53] the AI has gone, oh, what would a human [79:55] do? Oh yeah, let's do the churly petty [79:56] petty thing. And so that's that's what [79:59] it's done. So anyway, I just thought [80:01] that was really interesting and thanks [80:02] Mark for re rediscovering it. [80:05] >> Ju just as an aside, I've just put a a [80:07] link in the guest chat uh which probably [80:10] you've already linked to in past [80:11] episodes. Uh but if we're going to have [80:14] if we're going to have AI fight, so at [80:16] least we've now got the the option. [80:18] >> Okay, let me let me try. Is this Can I [80:20] put this on the screen? Is it safe? [80:22] >> Yeah, it's safe. [80:23] >> Right, hold on. Okay, we're all all [80:26] having a giggle, are we? Okay, here it [80:27] comes. [80:29] Oh, no. I haven't fe This is insane. I [80:33] haven't featured this. I actually saw [80:35] this the other day. Should we give the [80:37] viewers a few seconds of this because it [80:38] is kind of cool. Uh, China's dancing [80:41] robots. How worried should we be? So, [80:43] okay, imagine they getish [80:45] >> and uh, check it out. I'll just sort of [80:48] zoom in a bit. [80:50] >> I mean, what the heck? [80:54] Doesn't that make you I mean obviously [80:56] they're doing some sort of [80:57] pre-programmed repertoire, but [81:00] yeah. [81:02] What what are you are you are you [81:03] advocating putting these in the ring [81:04] with Remus? Is that why this is on the [81:06] screen? [81:07] >> No, I'm just suggesting that if you're [81:09] going to if we you know you seem to be [81:11] very keen on this idea that Paul Remus [81:13] is going to be put into a cage. [81:15] >> Yeah. Yeah. [81:16] >> Um and then you mentioned AI. I was [81:18] like, well, if we're going to put AI [81:19] against him [81:22] Yeah. They're coming for us sometime. [81:23] >> To be fair, I think he could with one [81:25] finger go like that and it would fall [81:26] over and just be sitting. [81:27] >> That's right. Yeah. Like a tortoise. So, [81:29] it'll be hilarious to watch. [81:31] >> Don't you think though that in like a [81:32] couple of years time, it's entirely [81:34] possible that we'll be looking back and [81:36] the the fawn will not be like that [81:39] anymore. It'll be, okay, now they really [81:40] can observe the finger coming towards [81:43] us, you know, them, and they'll just [81:45] snap that finger right off because [81:47] >> I think it's it's when you put the guns [81:49] on them that you need to start. But [81:51] we'll um we'll skip over that and go [81:52] with the [81:53] >> I was going to say we don't need more [81:54] violence in this particular episode. We [81:56] can all agree uh no guns for the robots. [81:59] Thank you very much. Right. I had a [82:00] couple of pieces which I'm not going to [82:02] have time to get to. Uh so let's see [82:05] where we get to now. Oh, this is another [82:08] piece that came from Matt. Um if you're [82:10] in the polyglots um ecosystem in the [82:13] WordPress space, this might be worth [82:15] looking at. It is AI related again as [82:17] everything seems to be and it's about [82:20] using uh AI to achieve the the things [82:24] which um traditionally the polyglots [82:26] team has used there. There's an awful uh [82:30] lot of well not a lot but there's a few [82:32] very incisive comments which come up but [82:35] um Matt is basically asking the question [82:37] what can we do with AI to make the whole [82:40] polyglots thing the translation um much [82:43] more straightforward and uh so if you're [82:45] into that go check it out right let's [82:48] have a little look um I've got one [82:50] product that I thought I'd mention this [82:52] week called review bird not come across [82:55] this before uh is for those of you that [82:58] are um using Woo Commerce stores and uh [83:01] I don't have a WooCommerce store and I [83:03] don't take product reviews so I don't [83:04] know how essential this kind of plug-in [83:06] would be but if you are and you want to [83:09] corral your product reviews go and check [83:11] out reviewbird.com [83:13] it allows you to manage all of that in [83:15] one place and given whenever I go online [83:17] and shop for things I do look at reviews [83:19] it's probably is quite an important part [83:21] of such a website right now that we're [83:24] um freaked ourselves out let's go to [83:26] this one to karma us all down, shall we? [83:28] This is Cloudflare. Their annual That's [83:32] not true. I think it's quarterly report. [83:34] Yeah, there you go. 2025 Q4 DDS threat [83:37] report. It's a It's a The title is [83:40] guaranteed to uh offer you sleep. But, [83:43] however, it says a record setting 31.4 [83:47] terabits per second attack caps a year [83:49] of impressive DDS DDoS assaults. [83:53] What didn't you know already from that? [83:55] But check it out. Like just some of [83:57] these numbers are absolutely ridiculous. [83:59] If you work for Cloudflare, God bless [84:01] you. I don't know how you guys manage [84:04] what you're doing, but uh here we go. Uh [84:06] the fourth quarter of 2025 was [84:08] characterized by an unprecedented [84:09] bombardment launched by the whatever the [84:12] heck that is. Arisuru Kim Wolf Botnet D. [84:16] >> They are famous botnet networks. [84:18] >> Oh, thank you. They have like uh they [84:20] have like over 200 last I checked 250 [84:24] million compromised devices handling. [84:27] >> Yeah. This was the thing that I was [84:29] saying earlier about the Google TV, you [84:31] know, the TV dongles and things like [84:32] that. Apparently, a lot of them are part [84:35] of this whole thing. Yeah. Um the [84:37] campaign targeted Cloudflare's customers [84:39] as well as Cloudflare as well as [84:40] Cloudflare's dashboard. [84:43] Now, I don't know if it's you, but I did [84:45] notice Cloudflare at the end of last [84:46] year going down a few times, like in in [84:49] a way that I'd never ever ever seen [84:51] Cloudflare go down before. And maybe [84:53] this explains it. Um, the hypervoluttric [84:57] HTTP DDoS attack succeeded. Just listen [85:01] to this number and try to imagine it for [85:03] a minute. Rates of 200 million requests [85:08] per [85:10] second. [85:11] I mean, what uh just weeks, that was [85:15] just weeks after a record-breaking 31.4 [85:18] terabits per second attack. Um and then, [85:22] you know, there's a bunch of charts [85:23] which ex which show fairly conclusively [85:26] that 2023 was lower than 2024. 2024 in [85:30] turn was significantly lower than 2025. [85:33] This graph looks logarithmic, by the [85:35] way, which means that by the year like [85:36] 2029, it's going to be infinite. [85:38] There'll be nothing else on the internet [85:40] apart from DDS attacks. It's just going [85:43] to go, "Oh, Tim." Yeah. Is that [85:45] plausible? [85:46] >> Um, yeah. I mean, [85:48] >> no. I I mean probably can answer this [85:52] much more clearly than I can, but yeah, [85:55] we we are we've already at the age where [85:57] we pretty much now [86:00] human traffic on the on the internet [86:04] that needs to be specified separate to [86:06] the web is such that we are now the [86:08] minority. Um and it will not be that [86:11] much longer before the web we are the [86:13] minority at which point yeah you you're [86:17] talking about automated systems talking [86:19] to automated systems. Um I I think these [86:22] numbers are are really interesting for [86:24] perhaps a completely different reason [86:25] and that's just how dominant Cloudflare [86:29] has become. [86:30] >> It's starting to worry me actually how [86:32] dominant Cloudflare is. [86:34] And as it happened, I think the times [86:36] that they were down, the big ones that [86:37] got in the news, each time was due to [86:40] conf literally uh conf errors rather [86:43] than actual I don't think they've gone [86:45] down from a compromise. But uh yeah they [86:49] are the just how dominant they've become [86:51] and how important they are to [86:52] infrastructure and uh whether that's a [86:55] good thing whether that's a bad thing uh [86:58] they are reaching that point where you [87:00] got to think about them and you you're [87:03] now in this position where you might as [87:06] you they've reached that tipping point [87:08] where up until recently there was like [87:10] well there's a little part way you might [87:12] you don't necessarily need to want to [87:14] use Cloudflare because you can use your [87:15] own but now If Cloudflare is down, so [87:19] much of the internet's down that so many [87:21] of the resources that you rely on for [87:23] your own website is probably down that [87:24] your website's probably down anyway, [87:27] >> right? Right. [87:28] >> So, you might as well use them. [87:31] >> The curious optimistic. [87:33] >> Yeah. But the the curious thing about [87:35] the Cloudflare thing is a lot of the [87:36] other platforms when they started to [87:38] gain dominance. I'll use Facebook as an [87:40] example [87:42] wrapped around that was a a growing [87:44] sense of like wow Facebook's like [87:47] there's a lot of wrong going on there. [87:50] There's a but I don't have that with [87:51] Cloudflare. I I don't get a sense of [87:53] there's a lot of wrong going on there. [87:55] It's just they're getting bigger and [87:57] bigger and bigger and so their failure [87:58] is just bad. I personally haven't read [88:01] enough. I've got no intuitions about [88:03] them being a, you know, and I'm doing [88:05] air quotes, a bad company doing [88:08] unethical things. Perhaps they are, but [88:10] I don't have that intuition. [88:12] >> They're a corporation at that scale. [88:14] There is always going to be some bad. [88:17] You can't be a corporation that size and [88:20] not [88:21] >> upset some person. [88:23] >> I have a solution, Tim. What we need to [88:24] do is put the is put Cloudflare [88:28] in a ring with Remus. it would lose [88:33] >> now that I think Remus can subscribe to [88:35] he's on team Cloudflare. I think maybe [88:38] they just shake hands and that would be [88:40] the end of it. Uh okay. Anyway, that is [88:42] really sanguin stuff. You know, the [88:44] numbers and the sort of the future that [88:46] we're painting ourselves into the corner [88:48] of. Kind of interesting, right? Some [88:50] more security stuff very very quickly. [88:52] CSS got hacked. [88:55] How is this possible? Go and read the [88:57] article. It's on Medium. It's kind of [88:58] interesting. It's complicated and [88:59] interesting and it's the way that the [89:01] Chrome browser uh pares. It's not C, you [89:05] know, you don't have to actually worry [89:06] about writing CSS apart from in some [89:09] really curious cases. But uh I've never [89:11] seen a vulnerability involving CSS [89:13] before. [89:14] >> It's just as well we have lots of [89:16] different browsers and we're not all [89:17] based on Chromium based systems, isn't [89:19] it? [89:19] >> Well, I know you are still So you still [89:23] using Firefox over there? Uh yes, except [89:26] for when doing broadcast to you and my [89:29] >> That's right. Because these platforms [89:30] won't allow it. Yeah, there's you and [89:32] your you and your seven other mates [89:34] still using I I love I love I really [89:38] want Mozilla to succeed. I listened to a [89:40] podcast episode with the the now new CEO [89:44] of Mozilla and he was trying to paint a [89:47] positive spin on their new AI [89:48] initiatives and I've got to say it [89:50] didn't wash too well with me. I I got [89:53] the bad feels from it. It was like sort [89:56] of like greenwashing what is a necessary [89:59] but horrible awkward evil that they're [90:02] having to deal with. Yeah. You in [90:04] agreement, Tim, a little bit. [90:05] >> Oh, yeah. Absolutely. [90:06] >> Okay. Great. Yeah. [90:08] >> Um [90:09] >> I think they released new feature to [90:11] disable all the the AI features. [90:15] >> Yes. Well, at least that's a thing, [90:16] right? You can go in there and switch [90:18] them all off, but ne necessarily they've [90:20] they're going to be enabling them by [90:22] default. But also, I just want a browser [90:25] clear of all that really. I want the [90:26] browser to just browse the internet [90:28] based upon where I click and nothing [90:30] more. [90:30] >> Obviously, there are the nice thing is [90:32] that there are just like there are [90:33] different Chromium browsers. There are [90:35] different Firefox variants out there and [90:37] people are working on [90:39] >> a good one. Libra Wolf, [90:41] that's very good one. If anyone who's [90:44] interested in privacy, Libra, what was [90:46] that one called? I [90:47] >> think called Libra Wolf. [90:49] >> Oh, okay. [90:50] >> As in like Libra Office, Libra free. [90:53] >> Yeah. Yeah. Perhaps from the Yeah. Libra [90:55] Wolf. It's a Libra Wolf variation. [90:59] >> Yep. I downloaded one not that long ago [91:02] which I didn't actually use because [91:03] there were significant things in the UI [91:05] which I didn't really like. But it was [91:06] curious because it was a it was a [91:08] Firefox clone but it was also trying to [91:11] enable Chrome and the Firefox apps uh uh [91:16] extension ecosystem. [91:18] >> Was that Waterfox? [91:20] >> It was I don't think it was. Um [91:22] >> there certainly the no AI pushing [91:25] heavily one. [91:26] >> Okay. But it was just the idea that [91:28] there are a few Chrome extensions that [91:30] I'm really familiar with and I would [91:32] love to be able to use and there's no [91:33] equivalent Firefox one. Um, but this [91:36] browser I thought might be able to do [91:37] both but it didn't manage it. There was [91:39] >> Firefox will install nearly all Chrome [91:42] extensions. [91:43] >> Okay. [91:44] >> Just as a complete aside if you because [91:46] they they they do share a commonality [91:48] for their APIs. [91:49] >> Okay. Okay. Uh so some comments coming [91:52] in about that. uh the entire Firefox [91:54] community. Oh, that's left. He's there. [91:57] Look, there there is the Firefox. [91:59] There's the UK Firefox community. [92:01] >> And do not worry, I am indeed up in [92:03] arms. [92:04] >> Yeah, is up in arms about that. I used [92:06] to be on their Sumo team. There's no [92:08] connection between the CEO and those [92:11] marketing the product. Okay. Um Right. [92:14] Okay. I think very very quickly we'll [92:16] just go through these. Uh I know time is [92:18] short and I don't want to hold the the [92:20] guests uh to stay for any length of time [92:23] longer than we said we would. Um if you [92:26] are using manage WP owned by GoDaddy, [92:29] they have now forged a relationship with [92:31] Patchack. I'll just mention that you can [92:33] go and read about that yourself. The I [92:36] don't know if this is the final round, [92:37] but if you're going to go to Word Camp [92:39] Asia or have intuitions that you might [92:41] like to go to Word Camp Asia, the fifth [92:42] round of speakers has been announced. [92:45] And then [92:47] uh we had the dancing robots already. D [92:50] I'm just going to drop this one and say [92:53] make of this what you will. This like [92:56] just look at that for a minute and pause [92:58] it. Just take in what you're seeing [93:02] and then tell me that like if you needed [93:05] proof of the existence of God or [93:06] something like that, this might be it. [93:09] Uh this is a fossil. like nature made [93:12] that and it was found by somebody on a [93:14] beach in and believe it or not they [93:16] found it on the island called Holy Holy [93:20] Island uh in Lindesvan and that's so [93:23] obviously somebody's mouth the human did [93:26] that there's no doubt about it or the [93:28] aliens the aliens left it there that's [93:30] probably what happened cuz look it's got [93:32] like the gums the pink gum bits and then [93:34] the teeth and then the Anyway apparently [93:37] not apparently it's a fossil it's on the [93:39] BBC website. Smiling fossil discovered [93:42] on Holy Island [93:44] drawn by teenager two weeks before with [93:47] crayon. [93:49] It doesn't actually say that, but I like [93:51] the idea. Okay, there we go. That's it. [93:53] We've run out of time. We've actually [93:54] overshot the amount of time that we [93:56] should be doing this show. Thank you uh [93:58] to you for joining in the comments. Uh [94:00] we await the obvious wrestling match [94:03] between Matt and Remus. That'll be [94:05] interesting to see. Maybe bring that [94:06] >> battle. [94:07] >> The rap battle. That's right. Yeah. No [94:10] violence. [94:11] >> No violence. The no violence rap mod [94:14] wrestling match. That's what I heard. [94:17] >> How about a break dance competition [94:19] instead? [94:20] >> Oh, what? [94:21] >> A break dance. A break dance [94:23] competition. [94:23] >> Yeah. Okay. Yeah. To to be honest with [94:26] you, what s what would satisfy me right [94:28] at this moment because I've been [94:29] watching so much of it is if they both [94:31] just had a go at curling. just do the [94:35] bit of curling and see who's the uh [94:37] >> who's the [94:38] >> I like the idea of the I feel I feel [94:40] feel that that still very much uh leans [94:42] more toward Remicus than it does Matt. [94:44] >> Yeah. Yeah. Okay. We keep present. Yeah. [94:46] Okay. Well, how about the my surname is [94:49] Cromwell. [94:50] >> Matt could hire that Australian uh one [94:53] person who competed in the last [94:54] Olympics. You remember her? [94:56] >> He could hire her to to Yeah. [95:00] watch [95:02] >> to do the curling cuz that would be the [95:05] curling break [95:07] >> can do all the moves and then finally [95:09] release it. I accidentally caught sight [95:12] of that Australian lady doing the [95:14] whatever that event is. Is it like [95:16] street dance or something like that? [95:17] >> It's broken dancing is what it was break [95:20] dancing. [95:20] >> It was on there was obviously a news [95:22] article. It was on in a shop that I was [95:24] in the other day and I watched it and it [95:25] really even though I know nothing about [95:28] that event, I have no chops about what's [95:30] credible and what's not. I was watching [95:32] it like, oh, it's so weird. It's so so [95:36] bad. Anyway, there we go. Thanks to our [95:39] team. Patricia says, appreciate it. [95:42] We'll be back this time next week, I [95:44] hope. Um, should be back this time next [95:48] week, but uh thank you. I'm just going [95:49] to go around the houses and say thank [95:50] you to Tim Nash. [95:52] >> There he is. Thank you to Mark [95:54] Benzakane. It's still like crazy o'lock [95:57] in the morning over there. Appreciate [95:58] that. [95:59] >> And also to Onar. Thank you so much for [96:02] joining us. Hopefully Onar, you'll come [96:03] back and join us on a less weird [96:06] episode. [96:06] >> They're always weird. They're always [96:08] weird. [96:09] >> They're not as weird. [96:10] >> Don't Don't fall for it. Every single [96:12] episode is weird. [96:13] >> Okay. All right. Okay. I probably should [96:15] rate that. [96:16] >> I I love it, though. This what makes [96:17] this fun, right? It's weird. So, [96:19] absolutely. [96:20] >> Yeah, I agree. [96:21] >> And uh we'll be back this time next [96:23] week. So, if the three of you want to [96:25] stick around and have a little bit of a [96:26] natural after this, I would welcome [96:28] that. But uh we will see you next week. [96:30] Thanks for joining us in the comments. [96:31] Appreciate it. Bye-bye. Bye-bye.  [[3. Resources/Readwise/Full Document Contents/Articles/This Week in WordPress #367|Full Document Contents]] ## Document Notes --- tags: #wordpress #core-development #security #performance #ai #open-source author: Nathan Wrigley source: WP Builds reference: https://www.youtube.com/watch?v=VMzU4GhuuCI date created: 2025-02-24 14:00 --- # 🏷️ Title: This Week in WordPress #367 ## 🧰 Action Inbox - [H1] Test WordPress 7.0 Beta 1 on non-production site — collaborative editing features #wordpress - [H2] Consider open source AI models for text-to-speech accessibility — cost savings #accessibility - [H3] Implement password manager if not using one — security necessity #security ## 🧭 Table of Contents - [[#📝 Summary]] - [[#📖 Main Content]] - [[#🗒️ Notes]] - [[#📚 References]] - [[#🗣️ Paraphrase]] - [[#📰 Newsletter]] - [[#✅ Action Items]] - [[#🧪 Hypotheses]] - [[#📊 Methodology]] - [[#❓ Questions]] - [[#🔮 Future Research]] - [[#💡 Implications]] - [[#🧠 Learning Styles Assessment]] - [[#📊 Accessibility Assessment]] - [[#🌍 DEIB Assessment]] - [[#✅ Logical Validity]] - [[#🛡️ DARVO Analysis]] - [[#🎯 Conclusion]] ### 🔗 References - [WordPress 7.0 Beta 1](https://wordpress.org/news/) - [MainWP Year in Review](https://mainwp.com/) - [Automatic Blog - Server Centric Development](https://automatic.com/blog/) - [OOPSpam Annual Report](https://oopspam.com/) ## 📝 Summary - Weekly WordPress discussion covering WordPress 7.0 beta release, AI developments, security updates, and industry trends - Panel discussion highlights tension between JavaScript frameworks and server-side development approaches - Key security updates include BitWarden audit response and password manager recommendations - AI discussion focuses on open source alternatives and cost-benefit analysis for WordPress users - WordPress 7.0 beta brings collaborative editing and design improvements ## 📖 Main Content This episode covers the evolving WordPress landscape with particular attention to the upcoming WordPress 7.0 release featuring collaborative editing capabilities. The panel discusses industry trends toward reconsidering JavaScript-heavy development approaches, with Automattic advocating for more server-centric solutions. Security remains a key focus with discussions about password managers and spam prevention. The conversation explores the tension between proprietary and open-source AI models, considering cost, privacy, and performance implications for WordPress users. ## 🗒️ Notes #### ⏱️ [00:00](https://youtu.be/VMzU4GhuuCI?t=0) — [H1] Introduction and Panel Setup - Key idea: Weekly WordPress roundup with international panel - Evidence: "This Week in WordPress #367, Monday February 23rd" #### ⏱️ [02:30](https://youtu.be/VMzU4GhuuCI?t=150) — [H2] Weather Check-in and Community Building - Key idea: Community engagement through casual conversation - Evidence: "Minus 5 here... visibility nil, blizzard warning" #### ⏱️ [05:45](https://youtu.be/VMzU4GhuuCI?t=345) — [H1] Andy Bell's AI Principles Discussion - Key idea: Some agencies taking principled stance against AI usage - Evidence: "He refuses to work with AI companies and won't use AI tools" #### ⏱️ [08:20](https://youtu.be/VMzU4GhuuCI?t=500) — [H2] MainWP's 12-Year Anniversary Review - Key idea: Long-term WordPress business sustainability and team motivation - Evidence: "12 years in the WordPress space is quite a marker" #### ⏱️ [12:15](https://youtu.be/VMzU4GhuuCI?t=735) — [H1] Automattic's JavaScript Framework Critique - Key idea: WordPress ecosystem questioning JavaScript-heavy development - Evidence: "A modest server-centric development proposal" #### ⏱️ [18:40](https://youtu.be/VMzU4GhuuCI?t=1120) — [H2] BitWarden Security Audit Response - Key idea: Transparent handling of security research findings - Evidence: "Security through Transparency - ETH Zurich audits" #### ⏱️ [22:30](https://youtu.be/VMzU4GhuuCI?t=1350) — [H3] Open Source Text-to-Speech Models - Key idea: Lightweight AI models for WordPress accessibility features - Evidence: "Kitten TTS is just 25 megabytes with 8 voices" #### ⏱️ [26:45](https://youtu.be/VMzU4GhuuCI?t=1605) — [H2] Spam Attack Trends and Residential IPs - Key idea: Increasing attacks from compromised consumer devices - Evidence: "Attacks coming from residential IPs, smart devices compromised" #### ⏱️ [30:20](https://youtu.be/VMzU4GhuuCI?t=1820) — [H1] Open Source AI Models vs Proprietary - Key idea: Quality gap closing but adoption challenges remain - Evidence: "90% quality difference closed within 13 weeks" #### ⏱️ [35:50](https://youtu.be/VMzU4GhuuCI?t=2150) — [H2] Hardware Requirements for Local AI - Key idea: Performance tradeoffs between local and cloud AI - Evidence: "Takes 45-50 seconds locally vs 5 seconds with ChatGPT" #### ⏱️ [42:10](https://youtu.be/VMzU4GhuuCI?t=2530) — [H3] Internet Archive and AI Scraping Conflicts - Key idea: AI scraping threatens historical web preservation - Evidence: "Guardian, NY Times, Reddit blocking Internet Archive access" #### ⏱️ [45:30](https://youtu.be/VMzU4GhuuCI?t=2730) — [H1] WordPress 7.0 Beta 1 Release - Key idea: Collaborative editing features approaching production readiness - Evidence: "Collaborative editing which many thought would be a roadblock is coming" ## 📚 References - [This Week in WordPress Show](https://wpbuilds.com/live) - [WP Build Show Feeling Insecure](https://wpbuilds.com/) - [Kitten TTS GitHub](https://github.com/kittenml/kitten-tts) - [Open Router](https://openrouter.ai/) - [TechDirt - Preserving the Web](https://techdirt.com/) ## 🗣️ Paraphrase - WordPress 7.0 beta introduces collaborative editing capabilities that were previously considered technically challenging to implement - The WordPress ecosystem is reconsidering JavaScript-heavy development approaches in favor of server-side solutions for better performance - Security remains paramount with password managers being essential tools, despite recent audit findings being largely theoretical ## 📰 Newsletter This week brings exciting developments in WordPress core with the 7.0 beta showcasing collaborative editing features that promise to transform content creation workflows. The community is also engaging in important conversations about development approaches, with Automattic advocating for more server-centric solutions over JavaScript frameworks. For security-conscious users, the BitWarden audit response demonstrates the importance of transparency in security tools. Meanwhile, open source AI models are becoming viable alternatives for accessibility features like text-to-speech. Resources: [WordPress Beta Testing](https://wordpress.org/download/beta/), [MainWP Community](https://mainwp.com/), [Security Best Practices](https://developer.wordpress.org/security/) ## ✅ Action Items #### WordPress - Core: Test WordPress 7.0 beta collaborative editing features in staging environments - Core Performance: Monitor impact of collaborative editing on server resources and database performance - Accessibility: Evaluate open source text-to-speech models for improved content accessibility - Security: Update security recommendations to include password manager guidance - Documentation: Create guides for server-centric development approaches vs JavaScript frameworks #### Community - Contributors: Share experiences with WordPress 7.0 beta testing and collaborative editing workflows - Maintainers: Review JavaScript dependencies and consider server-side alternatives where appropriate - Sponsors: Support open source AI initiatives that benefit WordPress ecosystem #### OSPO - Security Management: Review implications of open source AI models for enterprise WordPress deployments - License Compliance: Assess licensing implications of various AI model integrations - Community Engagement: Foster discussions about sustainable development approaches ## 🧪 Hypotheses - WordPress 7.0's collaborative editing will drive increased adoption among content teams - Server-centric development approaches will gain traction as JavaScript fatigue increases - Open source AI models will become standard for WordPress accessibility features - Password manager adoption will accelerate following security audit transparency ### 📊 Methodology - Beta testing WordPress 7.0 collaborative editing across different hosting environments - Performance benchmarking of server-side vs JavaScript-heavy implementations - User experience studies of open source vs proprietary AI integrations ### ❓ Questions - How will collaborative editing impact WordPress hosting requirements and costs? - Will the server-centric development trend influence WordPress core development priorities? - What role should open source AI models play in WordPress accessibility standards? ### 🔮 Future Research - Long-term performance implications of collaborative editing at scale - Security implications of self-hosted AI models in WordPress environments - Community adoption patterns for server-centric development approaches ### 💡 Implications - Practice: WordPress developers may need to rebalance JavaScript vs server-side skills - Policy: AI model choices increasingly impact privacy and cost considerations - Security: Transparent security practices become competitive advantages - Education: Training needed for collaborative editing workflows and server-centric development - Funding: Open source AI models require sustainable funding mechanisms ### 🧠 Learning Styles Assessment Reading level estimate: Grade 12 #### 👀 Visual learners - Present: Live video format with multiple speakers and screen sharing - Improve: Code examples and architectural diagrams for technical concepts #### 🔊 Aural learners - Present: Podcast format with conversational discussion - Improve: Audio-only version with clear speaker identification #### 🗣️ Verbal learners - Present: Technical terms explained in context, analogies used - Improve: Glossary of WordPress-specific terminology #### 🧩 Kinesthetic learners - Present: Encouragement to test beta versions and experiment - Improve: Step-by-step setup guides for featured tools #### 🧠 Logical learners - Present: Performance comparisons and cost-benefit analysis - Improve: Decision trees for choosing between alternatives #### 🧑‍🤝‍🧑 Social learners - Present: Live chat interaction and community engagement - Improve: Discussion forums for continued conversation #### 🧘‍♀️ Solitary learners - Present: Podcast format allows for individual consumption - Improve: Self-assessment checklists for covered topics #### 🌿 Naturalistic learners - Present: Organic conversation flow and real-world examples - Improve: Ecosystem thinking about WordPress sustainability ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Video has spoken content suitable for audio-only consumption - Improve: Formal transcript with speaker identification and timestamps #### Keyboard navigation - Present: Standard YouTube player controls - Improve: Chapter markers for easy navigation between topics #### Contrast and readability - Present: Video player meets standard contrast requirements - Improve: High-contrast mode for live chat interface #### Text resizing and reflow - Present: Responsive video player - Improve: Adjustable text size for live comments #### Media and captions - Present: Live spoken content - Improve: Real-time captions and downloadable transcript #### Forms and interactions - Present: Comment forms with appropriate labels - Improve: Better error handling for failed comment submissions ### 🌍 DEIB Assessment #### Cultural breadth - Present: International panel with different perspectives - Improve: More diverse geographic representation #### Inclusive language - Present: Technical terms explained in accessible ways - Improve: Avoid cultural assumptions about weather and time zones #### Neurodiversity support - Present: Multiple format options (video, audio, text) - Improve: Chapter markers and topic timestamps #### Representation - Present: Technical experts from different backgrounds - Improve: More diverse speaker demographics #### Psychological safety - Present: Respectful discussion of different viewpoints - Improve: Clear community guidelines for live chat ### ✅ Logical Validity #### Strong reasoning - Deductive strength: Technical analysis of WordPress features based on beta testing - Inductive strength: Trend analysis from multiple data points across ecosystem - Causal reasoning: Clear connections between development choices and performance outcomes - Analogical strength: Effective comparisons between AI model types and capabilities #### Fallacies - Appeal to authority: Some reliance on "experts say" without detailed evidence - False dichotomy: Occasional presentation of JavaScript vs server-side as either/or choice #### Tone of voice - Overall tone: Conversational and accessible while maintaining technical accuracy ### 🛡️ DARVO Analysis - Deny: No concerning patterns of denial - Attack: No attacks on critics or opposing viewpoints - Reverse victim/offender: No victim/offender reversals detected - Emotional cues and rhetoric: Generally balanced and factual discussion ## 🎯 Conclusion This episode effectively covers the current state of WordPress development with particular attention to upcoming collaborative editing features and evolving development philosophies. The discussion of open source AI alternatives and security best practices provides practical value for WordPress professionals navigating rapidly changing technological landscapes. ## 🧱 Refactor Appendix ### 📚 Literature Split Pad Purpose: Stage CEE units that become Literature notes. ### [H1] WordPress collaborative editing transforms content workflows > [!quote] This Week in WordPress #367 by Nathan Wrigley via WP Builds. Published 2025-02-24. > "Collaborative editing which many thought would be a roadblock is coming" WordPress 7.0 beta demonstrates technical feasibility of real-time collaborative editing features. [view highlight](https://youtu.be/VMzU4GhuuCI?t=2730#H1) #xref #articleseed ID:rw_wordpress_collab ### [H2] Open source AI models closing quality gap rapidly > [!quote] This Week in WordPress #367 > "90% quality difference closed within 13 weeks" Open source alternatives achieving near-parity with proprietary models in short timeframes. [view highlight](https://youtu.be/VMzU4GhuuCI?t=1820#H2) #xref ID:rw_ai_quality_gap --- ### 🔖 Atomic Split Pad (Zettel Candidates) #### [Z1] Server-side development reduces JavaScript complexity - Applicable to WordPress sites prioritizing performance over interactivity - Reference to Automattic blog post on server-centric approaches - Links: [[WordPress-Performance]] [[JavaScript-Frameworks]] #### [Z2] Password manager adoption essential for security - Universal recommendation regardless of specific tool choice - Transparency in security audits builds user confidence - Links: [[WordPress-Security]] [[BitWarden-Audit]] --- ### ✏️ Fleeting Queue - [F1] Test WordPress 7.0 beta collaborative features — staging environment setup #fleeting #inbox - [F2] Research Kitten TTS integration for accessibility — 25MB model evaluation #fleeting #inbox - [F3] Review MainWP anniversary insights — 12-year sustainability factors #fleeting #inbox --- ## Full Transcript [02:00] Hello, good morning, good afternoon, [02:02] good evening wherever you are in the [02:04] world. And I can see some of you from [02:06] different parts of the world. It's very [02:07] nice. Lots of comments dropping in [02:09] largely about the weather which is [02:10] always interesting. Um, so let's just [02:13] sort of say what we're doing this week. [02:15] It is this week in WordPress number 367. [02:19] I feel that's the wrong number, but [02:21] let's go with that. That's what I wrote [02:22] down. 367. [02:24] It is Monday, the 23rd of February. [02:27] We'll be repackaging this as an episode. [02:28] It will come out into your podcast [02:30] player of choice tomorrow. Tomorrow [02:32] morning at the the highly unusual 7 a.m. [02:36] in the UK. I don't know why I picked [02:38] that, but I picked it and thought I'm [02:40] sticking to it. I'm being consistent. [02:41] So, it's gone out at 7:00 in the morning [02:44] in the UK for about eight years or so. [02:47] No reason, but at least it's there when [02:49] you wake up, depending on where you are. [02:51] The idea is to talk about WordPress. Um, [02:54] but increasingly we end up talking about [02:56] all sorts of different things. And I [02:57] think there'll be quite a lot of [02:58] security stuff, possibly AI. Well, I say [03:01] possibly AI stuff. It's mostly AI stuff [03:04] these days. There's there's no oxygen [03:07] left in the room. Um, but in order to [03:09] make that happen, I'm always joined by a [03:11] panel of um, experts, let's put it that [03:14] way. And, um, today I'm joined by these [03:17] three fine people. There, there, and [03:19] there. And first one over here is uh, [03:23] Sir Tim Nash. Hello Tim. How are you [03:25] doing? [03:26] >> Hello. I I just can't remember when I [03:28] got kned. You'd think it would be an [03:30] important thing in my [03:31] >> Yeah, honestly, this whole world kne [03:34] down guy with a sword coming at me, but [03:36] no. Budget cuts. Budget cuts. [03:38] >> Yeah. Yeah. Yeah. Yeah. The royal family [03:41] are having a tough time. [03:43] >> Yeah. [03:44] >> Wonder if they'll start charging. [03:46] Anyway, let's not get into that. That's [03:48] definitely off topic. Um, but how you [03:51] doing? You all right, Tim? [03:53] >> Ah, yeah. I'm here. [03:56] That's about as much enthusiasm as you [03:57] can reach the moment. [03:59] >> That's the level of enthusiasm we're [04:01] gonna get. I ask the the contributors to [04:04] give us a little potted bio. And so Tims [04:06] Tims has Tims has always excited me in [04:09] the past. Uh I never read it just in [04:11] case it's something that requires a bit [04:13] of inspiration at the start. So here it [04:14] comes. I genuinely haven't read it. [04:16] Right. Here it is. Tim scares people and [04:19] oh no. Tim scares people and no more so [04:22] than Nathan Wrigley for he fears the [04:24] intros. He will be sing will he be [04:26] singing, dancing an intro or just saying [04:29] hello? I don't know. Uh today though he [04:32] should not fear he should fear not as [04:33] the no intro is oh is this intro is [04:36] really tame. Thank the lord. Tim is a [04:39] professional doom speaker and security [04:42] consultant specializing in WordPress. [04:45] You can find about you can find out more [04:46] about him at timnash.co.uk [04:49] or listen to his semi-regular chat with [04:52] Nathan on the WP Build Show Feeling [04:54] Insecure. We should do that again. We we [04:57] should we haven't for a recording for a [04:59] long time a long time and uh yeah let's [05:01] get back on the wagon with that but I I [05:03] am so pleased that there's there was [05:06] nothing that I had to sing or do as a [05:08] sea shanty or any of that this week this [05:11] time. [05:12] >> Don't worry I'm back in a few weeks [05:14] time. [05:14] >> Yeah. Okay. All right. I'll hold my [05:17] breath. Uh but thank you for joining us. [05:19] I'm also joined by Mark Benzane. [05:22] Hello Mark. [05:23] >> Hello. [05:24] >> How you doing? It's always like a [05:26] pleasure to drag you out of your bed at [05:28] 3:00 in the morning or whatever it is. [05:30] >> It's it's it's the pleasure is all mine. [05:33] And I mean that 100% sincerely. [05:36] >> What time is it? [05:37] >> Uh right now it's actually uh it's at [05:41] six o'clock. But by the time I roll out [05:44] of bed and, you know, take my shower and [05:47] do all the things and make sure that my [05:49] setup is right, especially this weekend, [05:51] my my computer crapped out and I had to [05:53] totally wipe and reload everything. And [05:56] I'm like, [05:57] >> yeah. So, so I'm like scrambling around [05:59] all weekend trying to make sure that [06:01] everything will work for the show today [06:02] and cuz I'm not going to have an hour [06:05] ahead of time to make sure things work. [06:07] >> No, no, no, no. That would mean getting [06:09] up at 2:00 in the morning. Well, that's [06:11] what that would mean. Exactly. Yeah. [06:13] >> If you're watching the show, you too [06:14] will notice that Mark spends his entire [06:17] life inside of a box. [06:18] >> I did. [06:19] >> Black box that follows him around [06:20] wherever he goes. [06:21] >> But I think outside of it. I think [06:23] outside of my my physical self is inside [06:27] the box. [06:28] >> You picked up what I was putting down. [06:29] Well done, Mark. That was great. Uh [06:31] Mark's bio goes like this. Uh Mark is [06:34] the partnerships and community lead at [06:36] MainWP. And though a totally contrived, [06:41] sorry, and through a totally contrived [06:43] coincidence, the timing lines up with [06:45] MainWP's 12th anniversary. Oh, nice. [06:48] Nice. Birthday celebrations. 12 years in [06:50] the WordPress space is quite a marker [06:52] and it felt right to acknowledge it [06:54] whilst he's here with us. Mark has been [06:56] part of the ecosystem for ages, working [06:58] closely with agencies, developers, and [07:00] product teams who manage sites day in [07:02] day out. His work at MainWP centers on [07:04] partnerships, community and helping [07:06] shape tools that support people building [07:09] and maintaining WordPress sites at [07:12] scale. Thank you, Mark. Really [07:14] appreciate it. [07:15] >> Thank you for having me. [07:16] >> Yeah, you're very welcome. And joining [07:18] us for the first time is Onar Aliy. [07:21] Hello. How you doing? Have I said you're [07:22] >> Thanks for having me. [07:24] >> Yeah, thanks for having me. I'm very [07:26] happy to be here. [07:27] >> Yeah, you're very welcome. Um, as as [07:30] Onar, it's his first appearance. there's [07:31] this empty bio box and no doubt on the [07:34] second appearance it will be full up [07:36] with junk, you know, because that's what [07:38] we do. We just mess it up. But on this [07:39] first time around, it's very obvious [07:41] that it's going to be sensible. So, here [07:43] is uh Onar's very sensible bio. He's the [07:46] founder of OOP SPAM. I'm not entirely [07:49] sure how you pronounce that. He can [07:50] correct that for us in a moment. [07:51] Formerly cyber security researcher [07:53] focused on online abuse and spam. And [07:56] he's based in Minneapolis, Minnesota. [07:59] How do we say it, Onar? [08:02] when you say we say oops ma'am. [08:04] >> Okay. I did actually think that would be [08:06] what it was but I didn't want to Yeah. [08:08] Okay. [08:09] >> I started a long time ago and now I'm [08:12] sinking back. Probably wasn't the best [08:14] name but [08:15] >> Yeah. [08:16] >> Yeah. Well, as you're about to find out, [08:18] I called this thing WP Builds and we [08:20] basically talk about cats. We has spent [08:23] ages on an episode the other day talking [08:25] about alligators. So [08:26] >> Oh, that's [08:28] hold on. Uh we'll see how that goes. If [08:31] you are joining us and you want to make [08:33] the show a little bit more interesting [08:34] and lively um then please make some [08:37] comments. That would be absolutely [08:38] lovely. The best way to do that is to go [08:41] here. Um Mark is sporting the little [08:45] banner there. Go share this stream [08:47] wpbuilds.comlive. [08:50] If you go there, you've got um you got [08:52] like two options really. The the the [08:54] most common option I think is to hit the [08:56] box on the side if you're on a desktop. [08:58] It's Google. So, it's YouTube comments. [09:01] So, if you're logged into a Google [09:02] account somewhere, you can comment like [09:03] that. However, if you don't want to use [09:06] a Google account or you don't have a [09:07] Google account or you mistrust Google or [09:09] any other reason for that matter, then [09:12] there's a box inside the video player [09:14] itself top right. It says start chat or [09:18] live chat or something like that. Click [09:20] on that and you don't need to be logged [09:21] into anything. You can uh be entirely [09:23] anonymous. And so please share it with [09:26] your friends, relations, animals, [09:29] pets, poltergeists, whatever the heck [09:31] you've got going on over there. And it [09:33] looks like some people have jumped in. [09:35] So here we go. We often get the weather [09:37] forecast. [09:39] Oh dear. Uh, and in Boston, James Laauo [09:43] is joining us. And it's 30. Oh, that's [09:45] brutal. It's minus one degrees [09:46] centigrade. Birds, birds, animals, tree, [09:49] branches, people fly at a 90 degrees [09:52] angle. Visibility nil. Officials, [09:54] blizzards warning roulette remain in [09:56] place until 7:00 a.m. ESC. Good. Good [09:59] morning. Oh, that sounds bad. [10:01] >> Well, it's minus5 here. [10:03] >> Oh. Oh. Oh, that's not real. That's like [10:06] the temperature on Mars or something. [10:09] >> But at least the internet works. So, [10:12] >> electrons love it. They love all this [10:15] cold. It's perfectly. [10:17] >> Uh Courtney Robertson's joining us, but [10:19] she's saying hello to James. Uh, she's [10:22] given a comment about the weather there. [10:23] Just two inches here. I guess that's [10:25] snow she's talking about. Oh, David [10:28] Warmsley, my partner in crime, David [10:30] Warsley. He says, "Hello, first time [10:32] catching the beginning of this show. [10:34] What a birthday treat." I'm presuming [10:36] it's your birthday, David. And if it is [10:38] your birthday, happy 92nd birthday. It [10:42] must be that at least by now. Uh, [10:45] something along those lines. I don't [10:47] know how old David is. I think he's [10:48] older than me then. But happy birthday, [10:51] David, if it is your birthday. Uh, the [10:53] webinar is also joining us. Hello, [10:55] beautiful people. How you doing? And [10:57] gosh, there's quite a few, isn't there? [10:58] Lots of people saying hello to each [11:00] other. Oh, come on. Rob rub it in, won't [11:03] you? David's in Goa where it's 27 [11:08] 27° centigrade. All I can say is the [11:11] electrons hate it. [11:13] >> I imagine the humidity is horrible [11:15] though. [11:16] >> Yeah. Yeah. [11:16] >> Go on. Go on. Tell me the humidity. [11:18] >> It's his birthday. He's getting old and [11:20] the humidity is horrible. That's it. [11:22] That's how we're going to how we're [11:23] going to reconcile it. Uh Ross Win, [11:27] hello. He says he's got the day off so I [11:29] can be live here today. A you are [11:32] friends, not work. That's so nice. Thank [11:34] you for joining us, Ross. Appreciate it. [11:36] Uh 12 degrees centigrade here in the UK [11:38] and very British. Yeah, I got the same [11:41] 12 degrees centigrade and just greyness. [11:44] You know, it's what we get every single [11:45] day of the year. Any other comments [11:47] joining in? Who we got here? Marcus [11:49] Bernette. Sorry. He says, "I had to [11:51] cancel you, Nathan." Oh, yeah. [11:55] Luckily, Onar's here. You know, that's [11:57] all I have to say about that. Marcus [11:59] wasn't able to come because he had other [12:01] things. And so, Onar very graciously [12:03] jumped in on Friday. And uh that's [12:05] amazing. Um testing WPB to one. We'll [12:08] get to that in a moment. As Jabber, [12:14] he's on skating on thin ice, is he [12:16] James? [12:18] Do I jabber? I think I jabber. [12:20] >> It's cold there. It's very thick ice. [12:22] Trust me. [12:23] >> Yeah, that's right. I think uh I think [12:26] is there any other comments that anybody [12:27] can say that I ought to put on the [12:29] screen because I'm kind of lost in them [12:30] now. There's too many to come up. [12:32] Anybody quickly? [12:33] >> Well, Courtney Courtney asked, but I [12:35] said I responded already, but uh [12:37] Courtney asked if the reset of my [12:39] computer means that the lights will not [12:41] go out. Uh I cannot guarantee anything. [12:44] Um, [12:45] >> Mark Mark's in a black box already, but [12:48] at some point the lights are go out, so [12:49] it'll be entirely black. It's really [12:51] funny when it happens. If you sit really [12:54] still, Mark, for like five minutes and [12:56] say nothing, we can all enjoy your [12:57] lights going out. Uh, okay. So, anyway, [13:00] there we go. That's the long [13:01] introduction. Really appreciate you all [13:03] being here. Keep dropping in some [13:04] comments. So, so, so the the idea is to [13:06] talk about WordPress. This is us, by the [13:08] way. Bit of short short promotional [13:10] stuff at the beginning. If you like what [13:11] we're doing over here, put your name [13:12] into that box. This is wpbuilds.com. [13:15] No, don't put your name. Put your email [13:17] address into that box. We don't need [13:18] your name. Just want your email address. [13:20] And then we'll send you a couple of [13:22] emails each week. One when we put this [13:24] out as a podcast. That'll happen [13:25] tomorrow when I've edited this. Um, and [13:28] then we'll put an episode out every [13:30] Thursday as well, and I'll send you an [13:32] email about that. And this week, uh, I [13:35] was lucky enough to have the fabulous [13:38] Andy Bell. Uh, Andy Bell. I think in [13:41] years gone by Andy Bell was more of a [13:44] WordPresser than he is now. I think it's [13:46] been many years I think since he's used [13:48] WordPress uh in in anger if you know [13:50] what I mean. But I had a chat with uh [13:52] Andy because I saw him at a conference [13:54] last year which Tim Tim was there as [13:56] well. It was in London last year and uh [13:58] Andy was on stage and he wasn't really [14:00] talking about WordPress. He was talking [14:02] about sort of agency life and things [14:04] like that. But the one of the things [14:06] that caught my attention was his [14:08] position on AI. Now Andy is very [14:10] principled about AI and he's very very [14:13] forthright. He doesn't want anything to [14:15] do with it. He doesn't want anything he [14:17] doesn't want his agency touching it. So [14:19] the content that they produce, the code [14:21] that they write, the websites that they [14:23] launch is all humanmade. [14:27] As soon as those words came out of my [14:29] mouth, I realized the error. Uh it was [14:31] all made by his agency, not humanmade. [14:35] Forgive me. Um [14:37] but uh but um anyway but he also won't [14:41] won't work with AI companies. So if an [14:43] AI company comes to his agency and says [14:45] will you build us a website? The answer [14:47] is going to be no. So we explore his [14:49] fairly principled position here. And I [14:51] got to say the more he talks the more I [14:55] was nodding my head you know the [14:57] dominoes kind of started to fall for me [14:59] and uh he was able to encapsulate it [15:01] really well. So I don't suppose any of [15:03] you three have got anything to add about [15:04] that. This is my promo bit at the [15:06] beginning, but I've dropped the subject [15:07] of AI and his principle on it. [15:10] >> I feel on principle, I should say. Did [15:12] the word lite turn up in the [15:14] conversation? [15:14] >> Uh, it should have done, shouldn't it? [15:16] Yeah, probably. But I I know I know what [15:19] you mean. [15:20] >> I think his is not really captured [15:22] around the technology. He's obviously [15:24] got a notion that he can do whatever the [15:26] heck an a an AI can do when it comes to [15:29] design. He's got the skills. He doesn't [15:31] need the shortcuts. And so whilst it [15:34] might save him a bit of time and [15:36] probably some money, go just go and have [15:38] a listen to him. I I know what you mean [15:40] though. But um I found myself nodding is [15:43] all I can say. [15:44] >> Uh anybody else, Mark Onar, anything to [15:47] add about that before we move on. [15:49] >> I I respect anybody who has principles [15:51] and sticks to them. Yeah, [15:53] >> that's that's you know um and so I I [15:57] have a great amount of respect for what [15:58] he's saying and if he can and I think [16:00] it's always important to be able to do [16:02] what the AI is doing. You have to be [16:04] able to understand what it's doing so [16:07] that you can either fix it or or or not. [16:10] So if you choose not to do it, it means [16:12] you know what you're doing. Well, and [16:14] the the other interesting bit was that [16:16] as as the sort of the conversation um [16:19] developed, not only was I kind of [16:21] getting more and more into it, but I [16:23] could see where he was coming from. And [16:25] the bit that kind of made the principles [16:27] stick for me was the fact that it really [16:30] hurt his business in the year 2025. [16:34] Incredibly successful and it had hurt, [16:37] you know, that having this principle had [16:39] hurt his business without a doubt. He [16:41] was very open about that. But what he [16:44] had not done publicly is he hadn't said [16:47] out loud the bit about the AI. He' just [16:51] not taken work that required him to use [16:54] AI. He hadn't taken work from AI [16:57] companies and he hadn't used AI. So [17:00] anyway, the knock-on effect was as soon [17:02] as he started saying the quiet thing out [17:05] loud, I refuse to work with AI, suddenly [17:09] his roster got filled up again. So it [17:11] turned out that for a small subset of [17:14] people and he presumably is an early [17:16] mover that that rounded off his 2025 [17:20] really well. Now we'll have to see how [17:21] long that continues, but it turned out [17:23] that that principle resonated with a lot [17:25] of people who needed a website. Um but [17:29] we'll have to wait and see how that that [17:31] works over time. Okay. Anyway, there we [17:33] go. Anybody else? Before I move on, [17:36] please keep your eye on the comments, [17:37] you three, because um if there's [17:39] anything that crops up, there's too many [17:40] at the minute for me to keep my um my my [17:43] eyeballs on. Uh by the way, um he has a [17:47] u like a course platform called uh [17:50] Pikilly. It's p i ca l i l. Pilly with [17:55] the li after the dot. And um a lot of [17:59] the courses are about this kind of [18:00] thing, you know. So mindful design is [18:02] one of them. I just thought I'd mention [18:04] that in case anybody was keen to pick up [18:06] on what it is that he's doing. Go and [18:07] check it out. Right. Okay. So, moving on [18:10] to user submitted content or in this [18:12] case guest submitted content. Here we [18:13] go. MainWP. You brought this to our [18:16] attention. It looks like the year in [18:17] review. How did MainWP go, Mark? [18:21] >> 25. [18:22] >> Well, first of all, Dennis asked me to [18:24] write the year in review this year, [18:26] which is actually a pretty intimidating [18:28] task, I got to say. And and I was kind [18:30] of like when you go back and and look at [18:32] everything that happened, it kind of [18:34] blows your mind at blew my mind as to [18:37] how much uh the team just did last year. [18:41] And I I think about it once again in the [18:45] frame of this 12-year anniversary thing. [18:47] And I just am blown away constantly that [18:51] they are so motivated every single day [18:55] to just keep pushing and pushing and [18:57] pushing after 12 years. I mean, that [18:59] just is astounding. So, that's the big [19:02] takeaway for me from this. The rest of [19:04] it goes into details as to how many [19:06] things we pushed out and and where we're [19:09] going and and all of that. Uh, and I [19:11] won't get into all of that. I I would [19:13] appreciate it if people would read it. I [19:14] appreciate you letting me share this uh [19:17] with the audience and I hope people read [19:19] it. And uh and uh I don't want to take [19:22] up too much more time other than that, [19:24] but uh I just I I just can't talk about [19:29] the group of people that Dennis has put [19:31] together at MainWP. It's just amazing [19:34] group. [19:35] >> Nice. Nice. Okay. Well, probably a quick [19:38] Google. Everything that we mention in [19:40] this show will be in the show notes [19:42] tomorrow. So when the podcast audio [19:43] drops, there'll be a link to a blog post [19:46] which is basically a bunch of links. [19:48] Some of them are in the show. There's [19:49] ones that didn't make it into the show, [19:51] but all of the ones that we're going to [19:52] mention will be in there. So, this will [19:54] make it uh into that list. It'll [19:56] probably be under the community section [19:58] if you go there, or possibly the plug-in [20:00] section, depending where I [20:01] >> where I drop it. But go and have a look. [20:02] Uh or you could simply Google at this [20:04] moment, couldn't you? Year of momentum. [20:06] A year of momentum 2025 year in review. [20:09] Nice that you love your team as much as [20:11] that. That's brilliant. [20:12] >> I I feel so lucky. Honestly, I I I went [20:15] from being a, you know, a an owner of a [20:18] company and a founder of a company to [20:20] being in a company where I feel like I'm [20:22] a founder even though I'm not. So, it's [20:25] it's it's really I I feel very fortunate [20:28] to have landed where I did. [20:29] >> Yeah. Oh, that's lovely. I'm just going [20:31] to pause the entire show very quickly [20:33] because I just want to say [20:36] over a period of about three months now [20:38] this person has dropped into the show um [20:42] Gank and every single episode I have no [20:47] idea what his name is and every episode [20:49] he says what his name is and I say I'll [20:51] remember it for next time and I never [20:53] remember it. So I've actually written it [20:55] down on a piece of paper hoping they [20:57] drop in today. I think you should name [21:00] the show after him. Yes. Just so that [21:02] you have an episode that's like [21:04] >> Yeah. Well, he's joining us from [21:05] Amsterdam. And of course, thank you, [21:06] Brian, for uh Oh, see what I did there? [21:09] Uh, it's Gerard. Thank you, Gerard, for [21:13] joining us. I really appreciate it. And [21:14] I I appreciate me. Sorry, I don't [21:17] appreciate me. I appreciate the fact [21:18] that you're coping with the fact that I [21:20] relentlessly, you know, been unable to [21:22] remember who you are. I really [21:23] appreciate it. And um, nice mention for [21:26] you, Tim. I think it was Ross I saw in [21:28] the comments who said, "It's lovely to [21:29] see you." Uh, what did he say? He said, [21:32] "Oh, no, it was Reese W. Nice to see Tim [21:35] again." That's [21:36] >> a Thank you. [21:37] >> Yeah, that's right. The bearded club. [21:40] They you you stayed together, you guys. [21:44] Okay. Have we got any nice comments in [21:46] there that Oh, that's very nice. Ross, [21:48] he's sort of saying, "Go and listen to [21:49] the interview, I presume, about Andy." [21:51] Uh, and staying on that topic, he says [21:54] Andy also said, so this is the previous [21:55] thing with Andy Bell. Um, that he'd got [21:58] some business from a client that had [21:59] been burned. Oh, that's interesting. [22:01] Yeah. By an agency that had used AI. Um, [22:05] and we're looking for somebody that [22:06] explicitly did not, which was kind of [22:09] interesting. [22:12] >> Um, and then David Wsley joins in. [22:14] Primary Andy problem solves and writes [22:17] HTML and CSS better than LLM seems to be [22:20] able to manage. I don't think you could [22:21] deny that somebody of his capacity [22:24] that's very likely to be true, isn't it? [22:26] Okay. All right. Let's move on a little [22:29] bit and we'll go here. So, that was the [22:31] bit that Mark dropped. Now, we've got a [22:32] couple coming from Tim's side of things. [22:34] I didn't get a chance to read this one, [22:36] Tim. Sorry about that. Um, so tell us [22:39] why this one's in. It's from Automatic. [22:41] >> So, it's sort of slightly slightly links [22:45] into Andy. Uh so this is a uh blog post [22:49] by somebody from automatic. They're on [22:52] the performance team at Automatic. Not [22:55] it's not specifically WordPress related [22:58] directly, but um it's basically a blog [23:02] post arguing we use too much JavaScript. [23:06] >> We use too React is React is not a [23:10] library. It's a framework. We use these [23:12] giant frameworks. They're ridiculously [23:15] complicated. They are really bad for [23:19] performance. They very rarely do the [23:22] things that we actually want them to do [23:23] well and we keep use trying to use [23:26] JavaScript when actually we have this [23:28] perfectly good servers behind the scenes [23:30] that are full of processing power that [23:32] could be doing the job for us. that a [23:35] lot of the time we struggle with [23:37] performance on websites is because not [23:40] necessarily the servers slowing down, [23:43] but your client has got to make a [23:45] 100,000 different requests backwards and [23:47] forwards to the server via the [23:49] JavaScript Ajax requests and JavaScript [23:52] is pulling and doing a lot of issues [23:55] when it comes to performance. And it's [23:57] basically a little bit of a h maybe we [24:01] should perhaps not use as much [24:04] JavaScript or only use JavaScript when [24:07] we need to and use it for what it's [24:09] really good at. Um, which I found quite [24:13] interesting. So it's not a get rid of [24:15] Ajax, get rid of React, get rid of all [24:18] of this. It was just a maybe we should [24:21] rethink that your if you're jumping into [24:24] some a project and your first reaction [24:26] is to jump and get react maybe consider [24:29] if that's the right course. And I found [24:32] that fascinating partially because it [24:34] very much mirrors my own viewpoint. So [24:37] of course I'm biased towards this as I [24:39] like this makes total sense. Uh I think [24:41] there's a lot of people in the comments [24:43] who will find this approach like oh yeah [24:46] that's what I do. that should be what we [24:48] should do. [24:49] >> But also who it came from. [24:51] >> Yeah, [24:52] >> it's on the automatic blog, [24:54] >> right? [24:55] >> I I'd like to think that this was a [24:58] little bit of a a flag has gone up, but [25:00] maybe we have gone too far the wrong [25:02] way. Um, but we'll see. But it's a it [25:06] was it's a really good read. It's a [25:07] really wellthoughtout read that um for [25:11] me uh reflects a lot of my own [25:14] viewpoints. So, I was sort of I was I I [25:17] very much was reading this and nodding [25:20] my way along. So, I think [25:23] >> I can more or less guarantee that David [25:24] Warsley in the comments will be warming [25:26] his hands on on this. He shares the [25:30] exact same opinion as you. Uh, you know, [25:32] buy you something when it's not [25:34] necessary. But you're right, that was [25:36] the bit that I took. I read the the [25:38] title and I read that it had come from [25:39] automatic and then I got to the bit [25:42] where it was obviously taking its first [25:44] like little targeted aim toward React in [25:48] particular [25:49] >> and I did think this is curious. [25:51] >> It is he he does take quite a strong aim [25:54] at React but then does emphasize that [25:56] it's not specifically React. It's more [25:58] that React is [26:00] >> act and I think his argument basically [26:02] is you never load just one JavaScript [26:04] library. [26:06] It's part of the uh philosophy of [26:10] JavaScript development seems to be well [26:12] I can just reach for this and I can [26:13] reach for that [26:14] >> and so you do end up with behemoths very [26:17] quickly and without even realizing it [26:19] you're you're suddenly you've got huge [26:21] amount of dependencies there that you [26:23] weren't expecting and yeah it's as I say [26:25] it's a really good read uh even if you [26:28] take away from it that is like that's a [26:30] bit extreme I if it just gives you pause [26:32] for four that maybe I don't necessarily [26:35] really need to bring in that full [26:36] framework to do this oneline thing. Or [26:39] maybe it could be just that you take the [26:41] opinion, oh actually I don't need to use [26:44] I can rethink how I'm using my [26:46] JavaScript and make sure that I'm first [26:49] of all the server can do this on its [26:51] own. you're almost certainly getting [26:53] going to have a better experience in [26:54] terms of accessibility, probably a [26:56] better experience in terms of access um [26:59] things like SEO and almost certainly a [27:01] better experience from a security [27:02] perspective as well because if you're [27:04] simplifying things generally that makes [27:06] it more secure, not always, but [27:08] generally the bit that's kind of [27:11] interesting to me is I would imagine [27:13] that the automatic blog has some level [27:15] of editorial control. I imagine that you [27:18] know uh whoever wrote this article and [27:21] the name has now gone Sergio Gomez uh is [27:24] the automatician who wrote this. I [27:26] presume there's some level of [27:28] inspection. So it is it is kind of [27:30] interesting to me that this has been you [27:32] know this has been given the the green [27:34] light and I I wonder I just wonder if [27:37] this is a first [27:40] in a cascade of similar things in the [27:43] year 2026. I think maybe that's what [27:46] you're I don't know. Maybe if you're [27:47] wondering that as well, Tim. [27:49] >> Yes. Not sure. [27:50] >> Well, no. Hoping maybe. I know. I feel [27:53] like this could either be a one shot and [27:54] it's like, "Oh, that's nice." Or we [27:56] could see this happening more and more. [27:59] >> And there is a little bit of a case of [28:02] >> as we are getting more and more AI slop [28:06] and a lot of if you ask a AI to build [28:10] something, it will pull in all of these [28:12] libraries and things. I can see the the [28:15] Andes of this world and people going [28:18] actually I can [28:20] win this battle by doing less and [28:24] >> producing good quality simple solutions [28:27] rather than pulling in the 50 million [28:29] libraries. So I can see it being used [28:31] that way as well. So [28:33] >> thank you. Okay, that's really [28:34] interesting. Uh I didn't say the title [28:36] of the piece and I have I'll read it in [28:38] because if anybody's uh listening to [28:40] this primarily this is on the [28:41] automatic.com. So with 2T's blog and [28:44] it's called a modest server centric [28:47] development proposal. So you can go and [28:50] look for that. It was written on the [28:52] 17th of Feb this year. So go and check [28:55] that out. Right. And there was another [28:56] one coming from you. Sorry I apologize. [28:58] I didn't let Mark and uh Onar have their [29:00] go on this if they wish. Otherwise we'll [29:02] move on. [29:04] It's [29:04] >> above my pay grade. [29:06] >> Okay. Me too. [29:08] >> I think you can find markets below your [29:10] pay grade. [29:11] >> Oh is that what you think? Okay. Okay. [29:13] >> Uh, okay. Right. Here we go. Um, next [29:16] one then again from Tim. Um, Tim's thing [29:19] is obviously security. I use Bit Warden. [29:21] I I don't know if they have products [29:23] outside of the sort of password manager, [29:24] but that's the length and the depths [29:26] that I have from them. And obviously it [29:29] being a password manager, I have this [29:31] assumption that it's going to be secure. [29:34] So, what's this one? What are we looking [29:36] at here? So this is a a blog post from [29:39] Bit Warden themselves called Security [29:40] through Transparency. ETHZurich audits [29:43] Bit Warden and cryptography against [29:45] malicious server scenarios. uh ETH [29:49] Zurich actually audited three or four [29:51] different password managers and claimed [29:53] that there were some serious faults and [29:56] I got pinged a lot [30:00] like every friend I know who I said go [30:04] use this go use a password manager [30:06] pretty much together it's like they got [30:09] sent the same email sent me going oh my [30:12] god it's all really had bad and it's [30:13] insecure it's like no no no the the the [30:17] group managed to find if you compromise [30:20] the bit warden servers and the clients [30:23] in this scenario and this happens then [30:26] something bad might happen. [30:28] >> Oh okay that's quite a lot of this is [30:31] there were variations on the theme for [30:32] each one of the password managers. So um [30:35] the reason I chose this as a good uh [30:37] post is because I thought it was a [30:38] really well-handled response from Bit [30:41] Warden because they didn't say this is a [30:43] non-issue. I thought that was really [30:45] good of them because they could have [30:46] just come out of the gate going, "Uhhuh. [30:49] Uhhuh. No." But they didn't. They [30:51] actually said, "We've taken things from [30:53] this. We've learned things from the from [30:55] this paper. We agree it's hypothetical. [30:58] Um, but honestly, [31:02] let's go. You know, we we there isn't [31:05] actually a problem, but we've learned [31:06] stuff." But it also gets gives me the [31:08] opportunity to highlight that you should [31:10] be using a password manager. um that and [31:13] the one of the reasons that I wanted to [31:15] highlight Bit Warden is that Bit [31:16] Warden's um behind the scenes both the [31:20] client is open sourced and the um the [31:23] server infrastructure is open sourced [31:24] and you can run something like uh Vault [31:27] Warden on your own infrastructure as [31:29] well. So, [31:30] >> okay, [31:31] >> there if you were if you haven't yet [31:34] your password manager [31:36] >> and I'm I'm disappointed in you, but [31:39] there are options out there. Uh, [31:41] obviously, if you have a password [31:42] manager of your own choice, go with that [31:44] one. Please don't don't change because [31:46] of somebody saying that Bit Warden's [31:48] more interesting because it's [31:49] transparent because of open source. [31:51] >> But that's why I wanted to highlight [31:53] this. [31:53] >> Yeah. Thank you. [31:54] >> To put a lot of people to bed so they [31:55] can stop DMing me about [31:57] >> Yeah, stop stop DMing Tim. Um, but I'm [31:59] glad because this is the one that I [32:01] settled on a few I don't know two or [32:03] three years ago when Last Pass just sort [32:04] of seemed to misstep horribly, [32:06] especially from a PR point of view. Um, [32:09] I moved over and pushed everything over [32:11] to bit one. So, I'm happy um that you [32:14] have the uh you know that you got the [32:15] skinny on this. Uh, by the way, Tim [32:18] paused about about a minute ago for [32:20] about three seconds and he scowlled. [32:22] There was a moment where Tim Tim scout [32:24] it was a very effective scallally. He [32:25] was trying to encourage everybody to use [32:26] a password manager. So, unless you want [32:28] to be on the receiving end of Tim's [32:30] scowlling, then get yourself get [32:32] yourself a part of I feel there should [32:35] be scowling on your website, Tim. Uh, [32:38] for me, yeah, [32:39] >> I'm a nice person. I don't I I want to [32:41] have clients. [32:42] >> Okay. Scowlling at them before they sc [32:45] my clients. [32:45] >> I would go and get a password manager. [32:47] It's not worth it, everybody. The other [32:49] thing was Mark's lights went out. I [32:51] don't know if anybody noticed that, but [32:52] he scuttled away very quietly and got [32:55] his lights coming back on. But there we [32:57] go. Few comments in coming in. Um, [33:01] where was it again? So, going back to [33:03] the um the bit that we were talking [33:05] about a minute ago. So, Andy Bell and [33:07] what have you. Um, let me just put it [33:09] like that. David Warsy says, so in [33:12] regard to No, it wasn't. It was about [33:14] the automatic piece and React and what [33:16] have you. Um, David Warsy says, "I just [33:18] wish WordPress was more active in the [33:19] W3C rather than looking to the corporate [33:22] web all the time." Uh, there we go. And [33:26] uh apparently Tim, you're a sucker for a [33:28] clickbait title. [33:30] >> Yeah, I'm not quite sure which one. [33:32] >> Denny, I think Scowl. Tim. Scowl. Just [33:36] give him the Yeah. Yeah. David knows his [33:39] place now. He's No more comments from [33:41] you, David. I feel like the soup. What [33:44] was it? The soup thing in Seinfeld. No [33:46] soup for you. No comments for you, David [33:49] Wsley. You're you're in the doghouse. [33:50] Right. Moving on. uh the so there's [33:53] three that are coming up from ONAR and [33:55] the first one I don't know if the this [33:57] GitHub repo is something that you're [34:00] involved with personally but we're [34:02] looking at kitten TTS which is a a [34:05] subdirectory of kitten ML what is this [34:08] what is this ML so this is text to [34:11] speech model uh which people can host u [34:15] self host it's only like 25 megabytes [34:18] smallest one this is uh I'm a big fan of [34:22] open source models. Um and in term uh [34:25] that's great for privacy privacy fox [34:28] people and in general this model what [34:32] gives you is you know sometimes we go [34:34] articles and there is listen to this [34:36] article and you can do this with this [34:40] model and without paying anyone you can [34:43] just host in your WordPress website in [34:45] your hosting hopefully you are not using [34:47] share hosting um so and so this will be [34:51] enough resources to run this but still [34:54] um perhaps it's impossible to sh run it [34:57] on share hosting and 25 megabytes quite [35:00] small and does a really good job um [35:03] converting your articles into tech and a [35:06] speech and comes with I believe eight uh [35:10] voices. Yeah, I just want to I I think [35:13] this team done a really great great job [35:16] getting this out and supporting open [35:19] source models and yeah so it says on the [35:24] the GitHub repo kitten TTS is an open [35:26] source realistic texttospech model with [35:28] just 15 million parameters. That sounds [35:31] like a lot to me but presumably it's [35:33] not. Um designed for lightweight [35:35] deployment and high quality voice [35:37] synthesis. I confess I haven't used [35:39] anything like this largely because every [35:40] single thing that I do is actually me [35:42] speaking. But I do see this more and [35:45] more on websites wherever I tend to [35:47] roam. There's quite a lot of here's the [35:48] text and then here's the accompanying [35:51] usually AI generated I don't know audio [35:53] file which is almost like a just almost [35:56] like somebody reading it out. I um I [35:59] find them quite hard to listen to most [36:01] of the time because they don't sound [36:02] real, but maybe this is [36:04] >> there's there's some examples down [36:06] there. If you scroll, you will see [36:08] there's a they put some examples where [36:10] they read the text. [36:13] >> Yeah, there's all different kind of [36:14] voices you can choose and experiment. I [36:16] think what's great about it like you [36:18] don't have to rely on third parties and [36:21] and also I'm uh in terms of [36:24] accessibility cases like you people can [36:26] just listen to article [36:29] and yeah okay then I will I will link to [36:33] it in the show notes if you have a need [36:35] to put you know speech on your blog [36:37] posts and things like that inside your [36:39] WordPress blog this will do that for you [36:41] by the looks of things. Uh, I'll just [36:44] Yeah, the the URL is not that difficult [36:45] to read, so I will just read it out. I [36:47] don't normally from GitHub because it's [36:48] hard usually. So, it's [36:49] github.com/kittenml/kitten [36:53] tts. So, that was the first one. [36:56] >> I was so disappointed it wasn't what I [36:58] thought it was. I have never clicked a [37:00] link on quicker expecting it to be [37:03] realistic meows. [37:05] >> Oh, I see. Okay. [37:06] >> I was I was I was all in for this. There [37:09] was going to be this realistic set of [37:11] meows coming out of it and I could [37:12] program my me some me some cat purring. [37:16] >> No, [37:17] >> there's a whole there's a whole [37:19] subculture of people. [37:20] >> I want a large language model that has [37:22] been trained on meows and purr [37:25] >> translated into cat. Yeah. Okay, that's [37:30] okay. So, there's the first one. Uh [37:32] second one from uh Onar is well I [37:36] presume this is your own because it's [37:37] the spam. You've got an annual report as [37:40] well. You've got some key findings here [37:42] but just just sort of run us through the [37:44] the highlight items of [37:46] >> Yeah. Yeah. I'm not going to take spend [37:48] too long on you on this. I just uh this [37:52] is usually we do every year where we [37:55] take some sample of data and look at the [37:58] data and this is a very actually related [38:01] to what Cloudflare um posted recently [38:04] about their um denial of service attack. [38:09] uh in recently we've seen a lot of um [38:12] attacks are coming from residential IPs [38:14] meaning like people they don't know [38:17] their devices are compromised like they [38:19] have smart boxes at home or they have [38:22] like their computers they their smart [38:25] devices they all like spreading u [38:29] starting spam campaigns or doing card [38:31] testing attack people are income are [38:34] very aware of this issue card testing [38:36] attack where people steal other people's [38:38] those cards, credit cards and they test [38:41] it on your WooCommerce website [38:43] and yeah this is about report about that [38:47] and yeah in general um awareness about [38:50] residential proxies [38:52] and yeah we seen again this year we saw [38:55] more attacks in WooCommerce and and also [39:00] WordPress is more popular so it has it [39:02] has seen more spam campaigns targeting [39:05] WordPress [39:07] um Yeah. Um that's all I Yeah. I [39:11] encourage people to read about it and [39:13] there is some um also free ways free [39:16] tools how they can protect themsel from [39:19] the attacks and against [39:23] >> yeah I will link to that again in the uh [39:26] in the show notes. Let me just see the [39:27] UR the URL is easy enough to say so I'll [39:29] read that in. [39:30] >> I think quite a lot of people who [39:31] consume this podcast they do not um they [39:34] don't read the you know the associated [39:36] posts. So, it's quite good if I can read [39:38] the URL. So, it's uh oops spam. So, [39:41] oopspam.com2025-spam- [39:47] report. So, you can go and check that [39:49] out and there are some key findings [39:50] there. We'll get on to a piece a bit [39:52] later if there's time from Cloudflare [39:54] and one of the things that they [39:55] particularly mentioned was the um the [39:57] use of uh smart TV [40:00] um hardware as one of the things which [40:02] has sort of ramped up I think over the [40:05] last quarter. We have a like a little [40:07] thing which we plug in the back of the [40:08] TV. It's I think it's Google TV [40:11] something like that. [40:11] >> Yeah. Also fire. [40:13] >> Yeah. [40:16] >> Okay. Right. In which case I think [40:18] you've got a third one as well. This [40:20] one. Um so this one again from Onar. [40:24] This one's called Well, it's on the MI I [40:26] don't know if it's MIT Sloan or MITS. I [40:29] don't really know. I think it's part of [40:30] the MIT uh infrastructure. It's the [40:32] Sloan school which has something to do [40:34] with I guess management or something [40:36] like that. Uh the title is open and [40:38] sorry AI open models have benefits so [40:41] why aren't they used more widely? Is the [40:44] answer to that because they're not as [40:45] good? Is that basically what it is at [40:48] the moment? Would that sort of [40:50] >> article also mentions about that um [40:52] there's usually 90% [40:56] quality difference. So usually and even [40:59] like uh even open source models they [41:01] close that gap very like in within 13 [41:04] weeks [41:06] >> and uh these models in the beginning [41:08] they were only like u open AI and all [41:11] the closed sources models so people more [41:15] got used to those models and people [41:17] started integrating to into their system [41:20] but then later there's more smarter [41:22] there's more smarter and equally yeah [41:26] high quality models also came out open [41:28] source they are yeah I think now it's [41:32] really expensive for companies to switch [41:34] I think that's one reason um people [41:37] don't want to deal with it but there is [41:41] and some people are concerned about open [41:44] source models because of privacy but [41:46] then there is open source models you can [41:48] host it in your own infrastructure if [41:50] you're a company and that's uh will be [41:52] told privacy and you own all your data [41:55] >> yeah and I think another concern they [41:58] raised there is um top open source uh [42:02] models now currently all made in China [42:06] >> and so they all come from yeah and [42:09] that's another privacy concern but again [42:11] you can host this uh this all the open [42:15] models locally and without any internet [42:18] connection [42:20] so that will solves the privacy issue. [42:23] Yeah. Uh in general the article goes [42:25] through why people should switch and [42:27] they should consider it and of course [42:29] there's a huge cost and yeah I um I have [42:34] this feeling that the year 2026 I'm just [42:37] taking this data from people out in the [42:39] wider ecosystem you know tech journalism [42:41] more broadly so not really WordPress [42:43] related and many of them seem to be [42:45] talking about the the real financial [42:48] upside down pyramid basically this sort [42:50] of really difficult to maintain [42:53] financial cial structure that a lot of [42:54] these AI companies are built upon. So, [42:57] you know, venture capital money, which [42:58] at some point the venture capitals will [42:59] wish to have back and and how the a the [43:02] AIs that many people have become [43:04] dependent upon will become significantly [43:06] more expensive for the exact same thing. [43:09] I do wonder as well g given the kind of [43:12] stuff that I would ever use AI for, [43:14] which might be literally just [43:15] summarizing a URL, you know, poke a URL [43:18] at it and say summarize that in a [43:20] hundred words. I I'd be I'm sure that [43:22] some of the open source ones would be [43:24] completely capable for me at that kind [43:27] of thing. Obviously, if you're trying to [43:28] do the cutting edge um coding and things [43:32] like that, then I'm sure that you do [43:33] need to reach into your pockets, but I [43:35] don't know, maybe open source will win [43:37] this in the end, but let's wait and see [43:38] if they're capable of using accessing [43:42] internet resources and using different [43:44] tools. Yeah, they much Yeah, they're [43:46] very capable nowadays. it just they [43:49] there's not as as much commercial [43:51] interest in open source models. So [43:53] there's not as much PR team behind them. [43:55] So I know [43:56] >> it's kind of kind of interesting how a [43:58] lot of the people in the open source [44:00] arena [44:01] um who advocate massively for open [44:04] source are quite happy to use these [44:06] proprietary technologies to uh you know [44:09] to build this that and the other thing [44:10] on top of WordPress. It's kind of [44:12] interesting how that whole [44:13] >> Yeah. You don't even have to like you [44:15] don't even have to host them yourself. [44:18] You can for example the open router um I [44:21] don't know you heard about this open [44:22] router is a infrastructure that hosts [44:24] all the different open-source [44:26] um products uh open source models and [44:30] you can use them through the companies [44:33] too and you get privacy benefits and [44:35] it's much much cheaper to use them to [44:38] for especially small uh simpler tasks. [44:42] So Ross, who uh who is in the comments, [44:45] sorry uh Tim, I'll just read Ross' [44:47] comment quickly. Um it surprises me says [44:48] Ross. Um the amount of the amount that [44:51] the WordPress project is talking about [44:52] AI so much but not talking about open [44:55] models. Sorry Tim, I think you were [44:56] trying to say [44:57] >> I was just going to put in a couple of [44:58] points. One is that there are two [45:01] different costs. open open source models [45:04] you have the cost to train the models [45:06] and the size that and infrastructure you [45:09] need to train the models and that is a [45:12] hardware problem that we just there [45:14] aren't companies going hello I'd like to [45:16] I have I have a data center over here [45:18] that I wish to train the models on I [45:20] will then hand that that open model back [45:22] into the real world and give it away for [45:25] free with the exception of unless that [45:28] is literally a company strategy or a n a [45:32] nation state strategy which is the [45:34] Chinese model process which is we're [45:36] going to throw this out as a disruptor [45:39] and that's [45:40] >> oh [45:41] >> really useful for us you know people who [45:43] are using the open source model they're [45:44] not doing it because it's an for open [45:47] source in the nice way we think of [45:50] they're doing it to disrupt against the [45:52] their larger competitors um and to stop [45:55] necessarily proprietary models taking [45:58] single hold if you do not have someone [46:00] who has enough hacking to give you the [46:03] data center to train these models. It's [46:05] really difficult. Then we have using the [46:08] models themselves. They still require on [46:12] the whole significant hardware. You were [46:14] saying, "Oh, can I summarize something?" [46:15] Absolutely. [46:16] >> Yeah. [46:17] >> Um the uh I have a a Quen model on my [46:21] laptop. Absolutely perfectly fine for [46:23] summarizing. takes about if I f if I [46:27] pointed at a uh an article it takes [46:30] about 45 to 50 seconds to come back to [46:32] me with a summarization [46:34] >> right right which is [46:35] >> I if I pass that to chat pt it come back [46:38] in 5 seconds [46:39] >> yeah okay there's the difference right [46:40] that's the [46:41] >> quality would be there that's fine for [46:44] one article scaled once on my machine [46:48] the problem comes when you want to do [46:50] heavy lots of workload so let's say I'm [46:53] doing I want to audit some code [46:56] um with a open- source models as we have [46:59] now let's even if it's a relatively [47:02] small code base [47:04] with my current hardware which I don't [47:07] have I'm not a massive company with lots [47:09] of hardware but we just hardware that I [47:11] have currently I can do it it just takes [47:14] a really really long time [47:16] >> whereas if I threw that same code base [47:19] at claude code it would be done in a [47:22] fraction of the time at that extreme [47:24] cost. I can throw my code base via [47:27] something like open router, but actually [47:30] open router is quite expensive because I [47:33] have to make the if I'm making a per API [47:35] request. I'm I'm still paying for their [47:39] hardware at much more inflated rate [47:42] because they have to make money. Open [47:44] routters is not a charity. They make [47:46] money on those requests through. So we [47:48] haven't yet reached a point where we [47:50] have models that are small enough to run [47:52] locally at performance that's required [47:54] to make the open source a viable model [47:57] and what we don't have yet is model a [48:00] way of training models that's small [48:02] enough that they or we can distribute [48:05] well enough um we we we will I'm sure [48:10] see in the next sort of 6 months to 12 [48:12] months um seti style setups if you [48:16] remember the oldi Yeah. Yeah. Hive mind [48:19] kind of thing. [48:20] >> Yeah. I'm sure that we will that's the [48:22] route that open source will ultimately [48:24] go down to solve the distribution [48:27] problem at making that. But to do that [48:31] introduces [48:34] amount of complexity, [48:36] security implications and pollution of [48:39] training data. Also just the whole how [48:42] do we train these models on what do we [48:44] train these models? Um because one of [48:46] the things about the open source models [48:48] coming from China is that there's at [48:49] least allegations that they're trained [48:51] on other AIs um which and the quite [48:55] often the training methods for these [48:56] larger larger open source models isn't [48:59] as transparent as we'd like. We get the [49:01] end result as an open source model but [49:03] how it got there is sort of like fuzzy [49:06] on the way through. Um that is the title [49:10] of today's podcast but there's no way [49:12] that I can sound like that. I want to [49:14] say the word fuzzy. [49:20] >> Make it make it in a really small font. [49:22] >> Tiny font. Yeah, tiny. [49:24] >> Tiny font. [49:25] >> Um, okay. So, that's really interesting. [49:27] Ross has got just sort of to we put that [49:30] comment up, but I'll just put it up one [49:32] more time. Um, that was the context for [49:35] Tim's reply. That is so interesting. If [49:37] you've got this intuition that some sort [49:39] of federated version of AI like uh like [49:44] the SETI thing that you mentioned where [49:45] people are sort of I don't know letting [49:48] their CPUs in idle time do some grunt [49:51] work in the background and and [49:53] contributing to a [49:54] >> probably better than turning it into a [49:56] crypto miner which is what those CPUs [49:59] are currently being used whether they [50:00] wanted to or not. [50:02] >> Yeah, I couldn't agree more. Um, I [50:04] however don't have access to a data [50:06] center, but just you know because it's [50:08] good news. I do have access to loads of [50:10] grapes and um and I'm going to eat these [50:14] grapes. My ambition during the course of [50:15] this podcast is to finish all the grapes [50:18] and uh and then really need a week. Just [50:21] to tie it back into security for a [50:23] second that the uh one thing we are [50:24] seeing more of is that um bad actors [50:27] when they are compromising servers are [50:29] doing so to get the CPU usage for AI [50:34] stuff for writing out spam emails and [50:36] blahy blah. Um not to the same it's only [50:41] in small pockets at the moment but I'm [50:42] sure we're going to start seeing. So if [50:44] you want to look for where we're seeing [50:47] uh distributed uh AI at the moment, it [50:51] is actually being used more as on the [50:54] aggression side. It is being used by bad [50:55] actors, not necessarily [50:58] for the benefit of mankind in general. [51:00] >> Who couldn't have seen that coming? I [51:02] know you're shocked. [51:03] >> Yeah, what a shame. I'm just going to [51:05] move I'm going to flick to something [51:07] that I was going to do at the end [51:08] because you were talking about um the [51:10] web being scraped and the way that you [51:12] know the open source models might scrape [51:14] it just as an aside. I was going to [51:17] leave this till the end. This is over on [51:18] the tech dirt blog. It's called [51:21] preserving the web is not the problem. [51:23] Losing it is. And this is just a curious [51:26] byproduct of that scraping that um [51:29] institution the institution which we've [51:32] all come to know and love the way back [51:34] machine. So the internet archives way [51:35] back machine is really caught at the [51:39] minute as to how best to deal with AI [51:42] agents scraping it because um it's being [51:46] used largely by AI agents scraping it [51:50] and a lot of the content for example [51:51] from institutions like it mentions a few [51:54] the Guardian the New York Times and [51:57] Reddit [51:58] um are now blocking the access to the [52:01] internet archive to scrape future [52:03] content. content, maybe past content as [52:05] well. I don't really know how that would [52:06] work, but future content. Um, because [52:10] there it's a way for the AI agents to [52:13] sort of get round their content [52:14] restrictions and what have you. And so [52:17] that would be a shame. That would be a [52:18] real shame. I kind of I didn't foresee [52:20] that one. You know, like I said a minute [52:22] ago, you could foresee the AI um, you [52:25] know, the miscreants misusing AI to [52:28] write spam emails, but for some reason, [52:30] I didn't see that one coming. And that [52:32] just seems like a a real shame if [52:34] something like that was to happen. Mike [52:36] Maznik, I think, wrote this. No, it [52:38] wasn't actually. It was Mark Graham. [52:39] It's not very long. It's pretty quick to [52:40] read. So, go and check it out. But it is [52:42] a bit of a shame. [52:44] >> The Internet Archive has for has its own [52:46] set of controversies based on scraping [52:50] when it may or may not be allowed to. [52:51] So, there is something ironic about [52:53] this. [52:53] >> Yeah. Yeah. Yeah. That's right. It's uh [52:55] shouting fire in the theater or [52:57] something like that. Yeah. Yeah. Yeah. [52:58] Okay. Um Okay, let's move on. Let's do [53:00] some sort of WordPressy stuff. So, this [53:03] is the first one. So, WordPress 7.0. [53:06] Beta 1 is just around the corner. I [53:09] can't actually remember the day it [53:10] drops. Maybe it's dropped already. Can't [53:12] remember. Um, but anyway, it's coming. [53:15] Um, and so it's time we go through this [53:17] sort cycle of beta releases and then [53:20] finally the the actual release will come [53:22] fairly soon. Time to start playing. [53:24] Don't put it on a production site, but [53:26] there is some fa there is some pretty [53:28] good stuff in this release. If you want [53:30] to follow somebody who's been following [53:32] the collaborative editing side of [53:34] things, go follow Steve Burge from [53:37] Publish Press because he more or less [53:39] daily uh on X tweets about his latest [53:43] discovery um to do with this. But I'm [53:46] taking my guidance from him, but it [53:49] seems like the whole collaborative [53:50] editing, which many people thought was [53:52] going to be a real roadblock, you know, [53:54] a really difficult thing to pull off, it [53:55] feels like it's it's coming. Um and it's [53:58] going to drop. We'll see what kind of [54:00] hardware you need to make that work. Um, [54:02] and then there's another whole there's a [54:04] whole bunch of stuff around design [54:05] improvements. So, the video uh embed [54:08] block and the grid block and the icons [54:10] and the breadcrumbs and the headings [54:11] blocks have all had adaptations as have [54:14] the nav as has the navigation block. And [54:16] then there's a bunch of developery stuff [54:18] as well. You can read it all in this [54:20] piece. Um, it's called WordPress 7.0 [54:23] beta 1 and it's on the WordPress.org [54:25] news blog. So, go check that out and go [54:29] and have a play a bit like Steve does. [54:31] You three, anything about that before I [54:33] move on. [54:34] >> Don't run it on production. [54:36] >> Don't run it on production. Look, it [54:38] literally says it on the first line. [54:40] >> You know, unless you're me, in which [54:42] case you've updated your site to use it. [54:44] But for the rest of you, don't run it on [54:46] production. [54:46] >> Yeah, that's right. Just don't. Um, so [54:49] there we go. Oh, and look, there's [54:51] Steve. Ha, there he is. Uh, hi, Steve. [54:55] Um, I was just saying that everybody [54:57] should follow you if they want to learn [54:58] about the collaborative editing because [55:00] you every few days it seems you drop a [55:02] little nugget of something that you've [55:04] played about within WordPress 7 and uh, [55:06] yeah, I'm very grateful for you doing [55:08] that. Appreciate it. So anyway, go and [55:10] check that out. Just a quick mention, [55:13] uh, friend of this podcast, he's been on [55:15] it lots and lots of times, no doubt will [55:17] be in the future as well, Remus Dice. [55:19] Um, he's launching a new service. He's [55:22] he's had a service for a little while. [55:24] uh truer than north. But I I won't try [55:27] to pause this. It's a fairly lengthy [55:29] written article, but he is um sort of [55:31] pivoting what he does in the WordPress [55:33] space um and picking up the mantle that [55:36] Guten sorry Gildenberg put down the [55:39] project run by Jonathan Wald and Tammy [55:41] Listister um which they put on ice I [55:44] think just before Christmas something [55:46] like that. Um I don't think Remus is [55:49] literally directly picking it up. But as [55:51] it's gone out of existence, he can see [55:53] that there's a sort of um a hole in the [55:56] market there. So he's going to be trying [55:57] doing what they're doing. In other [55:58] words, communicating with people, trying [56:00] to figure out where they fit with their [56:02] products and services in the WordPress [56:04] ecosystem. So he's going to be like a [56:06] consultancy for that. He does however [56:09] have a direct rival because I spoke to [56:12] Matt Cromwell uh during the course of [56:14] this week. Um and he co-founder of [56:17] GiveWP, worked for Stellar WP for a long [56:20] time. So equally credible in terms of [56:22] their length and breadth of history in [56:24] the WordPress space. Um he's launched [56:27] something called Roots and Fruit which [56:29] looks like it's going to be exactly [56:32] similar very analogous to what Remus is [56:35] doing. So, if you need a bit of guidance [56:37] in your WordPress company's journey, [56:40] check out Rimkus' blog, but also if you [56:42] check out um the Tavern podcast that I [56:45] did with Matt Cromwell or go to [56:46] rootsandfruit.com, [56:48] you'll be able to see what they're both [56:50] up to. Anything about that? [56:51] >> In the pre-show, we were suggesting that [56:53] they should we should just set them up [56:55] in a ring and fight, but I've [56:56] subsequently thought that given that how [56:57] unfair that would be to Matt [56:59] >> Yeah. [56:59] >> that a rap battle might be in order [57:01] instead. [57:02] >> Oh. um rap battle. [57:05] >> Yeah. Could Could we not do both? Would [57:08] it be possible to have them both doing [57:09] that at simultaneously? A a sort of cage [57:13] fight meets rap battle all at the same [57:16] all at the same time. Uh I apologize [57:19] Matt. This wasn't my idea or Remus. It [57:21] was Tim's idea. I'm just saying. [57:25] >> But but Nathan, you did fully support it [57:27] before the show. So [57:28] >> No, no, no. I I don't remember any of [57:30] that. I can't I can't can't remember [57:32] that at all. No, no, no. Yes, it would [57:35] be fun. Um, and then Taco's here. So, I [57:38] know Taco knows Remus very well. Maybe [57:40] you could persuade him to get in a [57:43] cage fight with Mar. This is not [57:46] >> I don't think anyone I don't think [57:48] anyone wants to get in a cage fight with [57:51] Remus. I don't know. I [57:54] >> no there's there's probably like a [57:56] handful of people on earth who' be who [57:57] best roads I think there's not many and [58:01] they probably all live in like Norway or [58:03] something like that. [58:04] >> I'm going to go back to the rap battle. [58:07] >> Yeah. Okay. All right. So, uh Remus and [58:10] Matt, if you could uh just crack on with [58:12] that, organize the rap battle and then [58:14] we can report back on it. I [58:15] >> I feel like a major word camp that would [58:17] work nicely. [58:18] >> Oh, yeah. Yeah. see who's like the best [58:21] at advising clients. Look, I tried to do [58:24] some sort of rap gesture then. I sort of [58:26] did this. Don't know. Don't know. [58:29] >> I I feel like that would be the [58:31] cringiest thing ever. [58:33] >> Yeah, that's why it should be done, [58:35] Mark. That's exactly. [58:36] >> Absolutely. Absolutely. Be it would be [58:38] the most watched WordPress TV thing [58:41] ever, probably. [58:42] >> Jamie Mars can compare it. It could be [58:45] Oh, yeah. Oh, yeah. I think I think I [58:47] think we've cracked it. So yeah, Taco's [58:50] saying poor Matt. A not poor Remus, [58:53] though. I know she's saying, but Amber [58:56] likes it. She's got a giggle icon. And [58:58] frankly, if Amber likes it with a giggle [59:00] icon, it has to happen. That's all there [59:02] is to it. Keep it coming. This is This [59:04] is brilliant. We've lost Onar. I don't [59:06] know if we've lost Onar or Onar's camera [59:09] has sort of just died the death, but [59:11] we've we've lost Well, no, we haven't [59:13] lost him. He must still be there in some [59:14] capacity because I can see that he's [59:16] still on the screen. So Onar, if you can [59:18] hear me. [59:19] >> Sorry, I'm here. I just have a problem. [59:22] >> That's fine. If the camera's busticated, [59:24] no problem. Just stick with us for the [59:25] audio. [59:26] >> I'm listening. I'm here. [59:27] >> Yeah, that's great. Thank you. Okie [59:29] dokie. Let's move on a little bit and [59:32] have a look at this. Last week we [59:34] reported on the fact that um the WPCC, [59:38] now annoyingly there are two acronyms [59:41] that fit WPCC. [59:44] There's the WP [59:46] what is it? community credits or [59:48] something like that. And then there's [59:49] the WP community collective. I can't [59:51] remember. [59:51] >> I thought you were talking about [59:53] Wimbledon Wimble Wimbledon's public [59:55] cricket club [59:56] >> that Well, there's three now apparently. [59:59] Uh but that one doesn't often feature on [60:02] this show sadly. Uh but it should. [60:05] >> Sorry, it was it's Wimbledon and Putney [60:07] Commons was the is it along with the [60:10] parent council. Honestly, it's been too [60:13] long since we talked about them on this [60:15] podcast, but here we are. So, WPCC in [60:19] this case means the WP WP Community [60:23] Collective. Uh, last week we reported [60:25] that um one of the founder members and [60:27] board members um had stood down. Uh that [60:31] was Katie Adams Farrell as you can see [60:34] if you can see the screen. Um quick off [60:36] the news, the the position has been sort [60:39] of filled if you like. There's uh a [60:41] couple of people that have stepped up, [60:43] but Chris Reynolds is probably the most [60:45] notable one to mention. He has stepped [60:48] in as the interim uh president. The [60:52] other one is uh Jono Alderson [60:55] and uh yeah, so they're both going to be [60:57] acting as part of the WPCC uh endeavor. [61:00] I know Courtney's joining us in the [61:02] comments. She probably has a lot more uh [61:04] sense to say about this because she is [61:06] one of the founders as well. Um and [61:10] yeah, the endeavor of this project is to [61:12] is to you know sort of grow the [61:14] community in a fair, equitable, [61:17] transparent way. And uh I feel I've had [61:21] a few chats with uh Chris Reynolds. I [61:23] don't know that I've ever spoken to [61:24] Jonno Alderson, but my the only [61:27] experience that I've had with Jonno was [61:29] watching him speak um at a WordPress [61:32] event. And boy, he is a force of nature, [61:34] isn't he? He's really quite quite [61:36] something. he was able to speak on the [61:38] subject of SEO and it was really it's [61:40] powerful watching him. So we'll have to [61:42] see we have to wait and see. Tim's sort [61:43] of having a bit of a bit of a giggle [61:45] there. [61:45] >> I just like the idea that yes, Jonno is [61:48] a force of nature and we should leave it [61:49] there. [61:50] >> Yeah. Okay. Okay. [61:51] >> Just because out of context that sounds [61:53] amazing. I think Jonno would appreciate [61:54] that as a statement as well. [61:56] >> Yeah. So Chris has got a sort of fairly [61:59] muted like fairly political way of of [62:03] explaining what he wants to do as part [62:05] of his presidency. But um Jono is a [62:08] little bit more forthright. He says [62:12] WordPress so this gives you some idea of [62:14] maybe where Jonno wants to take it. Uh [62:17] WordPress runs on contribution but too [62:18] often the people and organizations doing [62:20] the work are expected to absorb the [62:22] cost. Alderson said, quoting him again, [62:24] "The WPCC is creating real structure [62:27] around sustainability, accountability, [62:29] and contributor well-being. I'm proud to [62:32] be joining the board and helping to push [62:34] WordPress towards a healthier, more [62:36] equitable future." This is on the [62:38] repository uh [62:41] blog, so you can go and check it out [62:43] over there. Uh anything from you three, [62:46] excuse me, about that? [62:47] >> Not about that, but can we just take a [62:49] moment? So, uh, you can't see this on [62:50] the podcast, but the amazing picture [62:52] that the repository found of Chris, [62:55] which is like you can't even see his [62:58] face. He's just [62:59] >> Which one is it? Is this him here? [63:01] >> Yeah. [63:02] >> Maybe he um maybe he assumes the camera. [63:05] He sort of lurks elsewhere. It's not [63:07] possible to find pictures of him because [63:10] whenever the camera comes out, he [63:11] disappears into a corner. [63:14] It looks like um like one of those [63:16] pictures that would be taken, you know, [63:17] you'd see on the front page of a of a [63:19] newspaper. I feel like that's right it [63:21] really well with the black and white and [63:23] it made me smile. It's like, yep, that's [63:25] that's the person. [63:27] >> Um let's see what Reese says. John Jonno [63:31] is a little bit more forthright is is [63:33] the understatement of the century. Okay. [63:36] Um I meant in that article in particular [63:38] because Chris said something and then [63:40] Jonno said something and John [63:41] >> Oh, no. No. life. Jono. [63:45] >> Yeah. Yeah. Uh, Taco says, "It's like [63:49] Jonno Fest." Jonno is absolutely bloody [63:52] brilliant, says Taco. Okay, [63:55] that's it. Ideal. Uh, D. And then I've [64:00] got some commentary here about [64:01] subscribing to my newsletter. [64:04] Uh, I don't know. I'll send me an email [64:06] and I'll figure it out. It's [64:07] adminwpbuilds.com. [64:09] He [64:10] >> was talking about the tavern, not you. [64:12] Oh, is he tavern? [64:13] >> Yeah, [64:14] >> I have no control over that, but send me [64:16] an email anyway at that email address [64:18] and I'll get it looked at by the team [64:20] that run the uh the Tavern website. [64:22] Okay, thank you very very much. Anyway, [64:24] there we go. Uh Courtney saying, [64:26] "Welcome to both Chris and Jono joining [64:28] the WPCC." [64:31] Uh oh, maybe it's one of Zach's. Zack [64:34] Zack, if you've ever met Zack, Zach's [64:36] like a total pro photographer. He takes [64:38] photographs of bands, mainly metal bands [64:42] for I think it's a side gig. It's not [64:44] like exactly what he does. He obviously [64:46] works in the tech space. Um, but side [64:49] gig. So maybe that is that one of yours. [64:51] Uh, Zach, did you actually take that [64:52] picture? That's interesting. Right. [64:56] Okay. There we go. That's that piece. [64:57] Right. Let's move on. I'm not [65:00] Unfortunately, I never gave this enough [65:02] time to really pause what was going on [65:03] here. I read it ever so briefly, thought [65:05] I'd understood it, and now that it's in [65:07] front of me, I'm not sure that I did. [65:08] So, I'll just I'll just say what I think [65:11] is going on here, and then I'll let [65:13] either the hive mind of the comments or [65:15] you three panelists tell me that I've [65:17] got it wrong. So, Matt Mullenweg took to [65:19] the make.wordpress.org Slack this week. [65:23] Is that right? Yeah. 18th of February. [65:26] Uh and said, "I've heard that w so word [65:28] camp Asia, Europe, and uh US have [65:32] separate Slack groups." Um but it would [65:36] be better to have everybody in the [65:37] main.org group. And then he says, "Let's [65:40] move everything over." He says, "I think [65:42] this was done because there was too much [65:44] friction in making new channels. Totally [65:45] fine to create public channels for big [65:47] events and private ones for planning, [65:49] etc. let's deputize a dozen people to be [65:52] able to spin up channels or admin them. [65:55] And so I don't know if this is a thing [65:57] which is now done or happening or what [66:01] the need was, but there was a there was [66:03] a few comments that came in saying it's [66:06] quite handy to have those things [66:08] separated in it. For example, at Word [66:10] Camp Europe, they like to keep it secret [66:12] where the um where the next Word Camp [66:14] Europe is. That's part of the whole [66:15] thing. you know, you show up and then [66:17] the final day you get to find out with [66:18] the with the expensively made video [66:21] where it's going next time. And [66:22] obviously, if everything was in a public [66:24] channel, that might not be the case. So, [66:26] I'm not entirely sure what the um what [66:29] the reasoning behind this is or whether [66:31] it'll happen. But the other thing which [66:33] is kind of interesting here, certainly [66:34] from my point of view, it says here, I'm [66:37] very open to water cooler or more [66:40] socials channels on Slack as well. We [66:42] should have some fun there as we work [66:44] together. So it's not just his endeavor [66:47] here is not just to have WordPress, you [66:49] know, code and events, but have other [66:51] silly things like a dog's channel. I [66:54] mean, it literally this is what's on the [66:56] page. A dog channel, a cat channel, a [66:59] games channel, a self-promotion channel, [67:01] and a Today I Learned channel. Now, [67:04] obviously, if you listen to this [67:05] podcast, you know I just like messing [67:07] about. So that that's music to my ears. [67:11] Let's have as many Let's have a hamster [67:13] channel. Let's begin there, shall we? [67:15] So, anyway, I don't know if this is big [67:17] news to you. If it's of no interest [67:19] whatsoever, I'll just leave it. If you [67:21] want to talk about it, we will. If not, [67:22] I'll just move on [67:25] whilst I'm eating grapes. [67:28] >> That's a big fat no. Nobody wants to [67:30] talk about that. Okay. Did any comments? [67:33] >> The only thing I have to say is Yeah, I [67:35] agree. We need to slack better. I'm not [67:37] going to say anything beyond that. So [67:38] the the headline is correct because [67:42] there's there's a lot of clutter in [67:45] Slack and and all the [67:48] social uh chat uh platforms. So I I [67:53] would like for there to be something [67:54] somewhat centralized, but I'm not sure [67:57] how it should be done. [67:58] >> Yeah. Yeah. But but just kind of [68:00] interesting. I didn't I suppose it it I [68:04] just don't know what the thought process [68:05] is behind it in terms of whether it's [68:06] needed or not. Um, but it's been mooted. [68:10] You can go and join in the comments that [68:12] there are quite, [68:13] >> you know, I mean there's not lots and [68:15] lots and lots of comments, but there's [68:16] quite a few comments. [68:18] >> I will be curious to know what they talk [68:19] about in cats and dogs. [68:21] >> Yeah. Yeah. [68:22] >> It's just cat videos. Cat dog videos. [68:27] >> Yeah. The cat cat channels got to be [68:29] more than dogs. [68:32] >> The dog channels exclusively cat for [68:34] cats. By the way, um, somebody's ears [68:37] were getting hot. You know, somebody was [68:40] could hear us over the internet. Remus [68:42] has joined in. Um, [68:45] >> yeah, I agree, Remus. I think that Matt [68:47] would hands down beat you in the [68:49] >> I was going to say, what's the [68:51] underlying What's What are you trying to [68:52] say there, Remus? What? I think you're [68:55] implying something, aren't you? You [68:57] know. Okay. What if What? What if? I'm [69:01] not even going to say it. It's so [69:05] what if Matt was allowed a small [69:07] rounders bat or something like that to [69:10] sort of offset the [69:12] >> You keep going to the violence rap [69:15] battle. That was the suggestion. [69:17] >> Never done violence before. [69:21] >> It's so funny though. I'm absolutely [69:23] loving I'm loving the rim because it's [69:25] even bothered to comment though. That's [69:27] absolutely brilliant. Um so anyway, [69:29] we'll see see if that happens. [69:32] I'm surprised he didn't put the word [69:33] fight in quotes while he was at it. [69:36] >> Yeah, that's right. Okay, let's see if [69:39] Matt Cromwell drops in and and offers it [69:42] offers it off. Okay, right. Let's get [69:44] into some sort of security stuff. [69:46] There's a few bits and pieces here. Um, [69:48] it's totally out of my pay grade this, [69:50] but I did think this was kind of [69:51] interesting. So, this is talking about [69:53] AI and security all in one go. So, [69:56] somebody that I followed on social media [69:58] for a little while, this is on [69:59] Masttodon, uh Kevin Bowmont, he says, [70:02] "Today in infosc job security news. I [70:06] was looking into an obvious [70:07] vulnerability introduced into a major [70:10] reb web framework today and it was [70:12] committed by username Claude on GitHub [70:16] vibe coded basically." So, I started [70:18] looking through Claude commits on [70:19] GitHub. There's over two million of them [70:22] and it's about 5% of o all opensource [70:26] code this month. As I looked through the [70:28] code, I saw the same class of [70:30] vulnerabilities being introduced over [70:32] and over again. Several a minute. [70:35] Honestly, I don't I don't think we need [70:37] to talk about that, but I just thought [70:38] that was an interesting thing to raise. [70:40] You know, just like what what what is [70:42] the future we're painting for ourselves [70:44] if if stuff like that is true? [70:46] >> Breaking news. AI was trained on bad [70:49] data. Bad input and bad data. [70:52] >> Oh, increases vulnerabilities. I'm [70:54] shocked. [70:54] >> No kidding. No kidding. Should we go to [70:56] the other interesting one that I found? [70:58] Where was that? Was it this one? No. Uh, [71:01] let me try. We'll come to that at the [71:03] end. Where's it gone? Where's it gone? [71:06] Where's it gone? [71:07] >> I think that one you were just on the [71:08] open the open AI has a open source has [71:12] an AI slot problem. [71:13] >> That was not the one I was thinking of, [71:15] but that is kind of interesting. No, [71:17] I've forgotten what the context of it [71:19] was. Maybe it was that one. No, I can't [71:22] remember. But there was one which sort [71:23] of jived perfectly with this one. Yeah. [71:26] Anyway, we'll come back to those in a [71:28] minute. But I just thought that was kind [71:29] of interesting. You know, we're painting [71:30] ourselves into a future where the robots [71:32] are now Oh, no. It was the one about the [71:34] robot. [71:35] >> Are you talking about the Sean the Sean [71:36] Hooper one? [71:37] >> Yeah, that's right. Which one was that? [71:39] Where? Let me see if I can find it. [71:41] >> It's uh when Rob When robots fight back. [71:45] >> Yeah, let me find it. You're right. [71:46] You've got the you've got the measure of [71:48] it. That's the one I'm [71:49] >> while you're looking for that. One of [71:51] the things that's interesting about that [71:52] is that these commits are being actually [71:54] submitted and are being submitted and [71:57] signed by Claude. So this is like hello [72:00] I we're not we're not even pretending [72:02] that there's someone looking over this [72:04] and have manually putting this in. This [72:06] is and this is just this is the AI has [72:09] submitted this. This is signed and [72:12] through. And there's part of me that's [72:14] really pleased that's h that's [72:15] happening. [72:17] because right now it's really difficult [72:19] if you get a piece of code in front of [72:21] you and you're like who who or what [72:25] coded this? Um so it would be nice to [72:29] actually see more AI code being uh [72:35] properly [72:37] um [72:39] offered by having showing who's actually [72:41] doing the authorship. Okay. [72:43] >> What the position of the human was [72:45] inside that. So actually I think we [72:47] absolutely know we are there not 5%. [72:51] It's going to be [72:53] >> masses more. It's clearly 5 5% of isn't [72:56] that's that's a that's not not the real [72:59] figure. It's going to be significantly [73:01] more. We in the space of just 12 months [73:04] the development space has completely [73:06] changed. If I'm not I reckon there's [73:08] significantly more than 5% has been of [73:11] just the WordPress co core code has been [73:14] committed. There's going to be way more [73:16] than that generated by AI in the last [73:18] commit set of commits for WordPress 7. [73:21] If you the real time collaborate if we [73:23] took one of those features the real-time [73:24] collaboration part and just looked at [73:27] how much of that was AI coded I suspect [73:30] that that is going to be quite high. [73:32] It's not going to be five. It's not [73:33] going to be 10. It's not even poss might [73:35] be even in 20 30s 40%. [73:38] But [73:40] you don't know that. [73:42] >> No, [73:42] >> because we have no easy way at the [73:44] moment of verifying other than looking [73:46] at it and going, "God, that's AI slop." [73:48] I'm not saying the real time stuff is AI [73:50] slop by the way that was probably [73:52] but that there is a but it gives you an [73:55] example. We at the moment we don't know. [73:57] So it would be good to have some sort of [73:59] ownership and some custody chain so we [74:02] can actually see what's generated that. [74:04] >> Okay. So you're saying that having some [74:06] huristics as to that portion was [74:09] definitely humans and that portion [74:11] wasn't having some reliable data on that [74:13] and people not lie. That's kind of [74:16] interesting. By the way, I never get to [74:18] say the word breaking on this podcast, [74:19] but I'm going to say breaking because [74:23] look, here it is. It's official. [74:27] Pat Cromwell has [74:30] >> he's joined in. [74:33] >> I I would also point out that Remicus' [74:35] reply of uh his his last reply [74:38] explaining that uh you know the dude [74:40] who's been doing Taekwondo for eight [74:42] years and kickboxing for free. And I'm [74:44] going to emphasize again in a rap [74:47] battle. [74:48] >> Okay. Okay. Where's the comment about [74:50] Taekwondo? Oh yeah. Okay. So just I mean [74:54] I'm not trying to [74:57] I'm really not but it just it's funny [74:59] right it's quite funny. Uh so Matt [75:01] Cromwell says my bets on the little guy [75:03] in that match and then Remus comes back [75:07] with mine are absolutely on the big dude [75:09] who's been doing Taekwondo for eight [75:11] years in kickboxing for three. Okay. I [75:14] think at this [75:15] >> I think the other thing that we have to [75:16] keep in mind here is Matt really is a [75:19] lover not a fighter. Okay. So there's [75:22] there's that, too. So [75:24] >> that's if if I um if I use that as the [75:26] title of this episode, it's going to be [75:28] misconstrued. Um if I use Matt's Matt's [75:32] really a lover, not a fight. I don't [75:33] think anybody's gonna pick off. [75:35] >> But it's true. Matt Matt, you know, Matt [75:38] would just go and give Remus a big hug [75:40] and that would be the end of it. [75:42] >> Okay, Taco, can I just say thank you? [75:44] Taco's now divorced us from all [75:47] responsibility. He's decided to take on [75:49] organizing it. That's great. We wash our [75:51] hands of the whole thing. It was a [75:53] terrible idea, but you Yeah, you can I [75:57] still submit a session? Yes, of course [75:59] you can. You do it. Let's uh let's get [76:01] them in a Let's get them in a ring. This [76:04] is hysterical. I love this. Uh and then [76:07] Matt, finally, rap battle. Okay, rap [76:09] battle it is to go with music majors and [76:11] native English speakers. Okay, I dare [76:14] you to use that title. [76:16] Lord, just [76:20] some episodes are easier than others. [76:22] Let's put it that way. I have to keep up [76:24] with the comments as well as trying to [76:26] make sense of everything that's coming [76:27] in. Okay, Elliot Richmond's joining us. [76:29] Hi, Elliot. Loving your blog, by the [76:32] way. It's very calming. Um, recommits [76:35] via AI. This is a very good reason for [76:37] AI governance ASAP. [76:40] How would that be done? Who would govern [76:43] AI? You'd have to have some sort of UN [76:45] for AI, wouldn't you? Because cross [76:47] >> EU EU has a bunch of laws around AI [76:51] safety and governance. [76:53] >> I don't know how that going. [76:55] >> I mean, the EU has a bunch of proposals [76:58] for laws around it, right? [77:01] >> But [77:01] >> I'm not sure they actually implemented [77:03] it or not, but [77:05] >> they they jumped on that train very [77:07] quickly. [77:08] >> We all know that governments are [77:09] excellent at do at understanding [77:11] technology usage. Oh yeah. [77:17] >> What team say about the you using AI [77:20] code comments? Uh I know that people [77:22] some companies I some I have some [77:24] friends in uh Shopify I think they also [77:27] publicly disclose that they use a lot of [77:30] AI comments but there's always human in [77:33] the loop that uh reviews all the pull [77:37] requests by um AI models and so you want [77:41] to make sure that uh all the codes are [77:44] align with your the standards your [77:47] internal standards and also there's no [77:50] vulnerabilities and stuff. [77:52] >> Yeah. I Yeah, I think if you just let [77:54] the AI to do pull requests and submit [77:57] the codes, then uh there will be a lot [77:59] of problems. [78:01] >> So, sorry, sorry. [78:03] >> Yeah, go ahead. Yeah, go ahead. [78:05] >> So, I didn't mean to interrupt. Sorry. [78:06] Um, but you said something which led to [78:09] the article which I was trying to find [78:11] which, uh, you got quite right, Mark. It [78:13] was the one and I've actually now Oh, [78:15] for goodness sake, I've now lost it. [78:16] There we go. It was the Shan Hooper one. [78:19] >> Uh, this came up this week. So Sean [78:21] Hooper um posted a piece called When the [78:24] Robots Fight Back. And I I'll just read [78:27] it because it's so short, but it's so [78:29] interesting. It says, "Today I saw [78:30] something I'd never seen before. A bot [78:32] taking revenge on a human. Matt's plot [78:35] lib is a popular open source Python code [78:37] library used to make visualizations. [78:39] Yesterday, a contributor named Krabby [78:42] Wrathburn submitted a bug fix. The pull [78:44] request was closed by Scott Shambbor [78:47] stating the project's rules that ban AI [78:50] submissions. Usually, this would be the [78:52] end of things, but no. This bot, which [78:55] runs the open claw AI, instead decided, [78:59] you can't make this up, decided to [79:02] create a blog post calling Scott out, [79:04] which it later retracted. [79:07] In my favorite part of this interaction, [79:09] the bot tells Scott, "Judge the code, [79:11] not the coder. Your prejudice is hurting [79:14] Matt plot lib. [79:16] What I mean what's happened to us? I [79:20] mean here we are talking about Matt [79:22] Cromwell and Remus having a fight. It's [79:24] this it's this it's completely hijacked [79:26] our minds. It's all it's all ahole. [79:30] I mean, seriously though, getting back [79:32] to that piece that isn't that just nuts? [79:35] Doesn't that like isn't every bit of [79:36] your senue going where how have we got [79:39] to this point where a a robot can can in [79:43] effect get the hump and then go off and [79:46] I mean obviously it's doing what humans [79:48] do, right? This is the sort of churish [79:50] petty behavior that humans exhibit and [79:53] the AI has gone, oh, what would a human [79:55] do? Oh yeah, let's do the churly petty [79:56] petty thing. And so that's that's what [79:59] it's done. So anyway, I just thought [80:01] that was really interesting and thanks [80:02] Mark for re rediscovering it. [80:05] >> Ju just as an aside, I've just put a a [80:07] link in the guest chat uh which probably [80:10] you've already linked to in past [80:11] episodes. Uh but if we're going to have [80:14] if we're going to have AI fight, so at [80:16] least we've now got the the option. [80:18] >> Okay, let me let me try. Is this Can I [80:20] put this on the screen? Is it safe? [80:22] >> Yeah, it's safe. [80:23] >> Right, hold on. Okay, we're all all [80:26] having a giggle, are we? Okay, here it [80:27] comes. [80:29] Oh, no. I haven't fe This is insane. I [80:33] haven't featured this. I actually saw [80:35] this the other day. Should we give the [80:37] viewers a few seconds of this because it [80:38] is kind of cool. Uh, China's dancing [80:41] robots. How worried should we be? So, [80:43] okay, imagine they getish [80:45] >> and uh, check it out. I'll just sort of [80:48] zoom in a bit. [80:50] >> I mean, what the heck? [80:54] Doesn't that make you I mean obviously [80:56] they're doing some sort of [80:57] pre-programmed repertoire, but [81:00] yeah. [81:02] What what are you are you are you [81:03] advocating putting these in the ring [81:04] with Remus? Is that why this is on the [81:06] screen? [81:07] >> No, I'm just suggesting that if you're [81:09] going to if we you know you seem to be [81:11] very keen on this idea that Paul Remus [81:13] is going to be put into a cage. [81:15] >> Yeah. Yeah. [81:16] >> Um and then you mentioned AI. I was [81:18] like, well, if we're going to put AI [81:19] against him [81:22] Yeah. They're coming for us sometime. [81:23] >> To be fair, I think he could with one [81:25] finger go like that and it would fall [81:26] over and just be sitting. [81:27] >> That's right. Yeah. Like a tortoise. So, [81:29] it'll be hilarious to watch. [81:31] >> Don't you think though that in like a [81:32] couple of years time, it's entirely [81:34] possible that we'll be looking back and [81:36] the the fawn will not be like that [81:39] anymore. It'll be, okay, now they really [81:40] can observe the finger coming towards [81:43] us, you know, them, and they'll just [81:45] snap that finger right off because [81:47] >> I think it's it's when you put the guns [81:49] on them that you need to start. But [81:51] we'll um we'll skip over that and go [81:52] with the [81:53] >> I was going to say we don't need more [81:54] violence in this particular episode. We [81:56] can all agree uh no guns for the robots. [81:59] Thank you very much. Right. I had a [82:00] couple of pieces which I'm not going to [82:02] have time to get to. Uh so let's see [82:05] where we get to now. Oh, this is another [82:08] piece that came from Matt. Um if you're [82:10] in the polyglots um ecosystem in the [82:13] WordPress space, this might be worth [82:15] looking at. It is AI related again as [82:17] everything seems to be and it's about [82:20] using uh AI to achieve the the things [82:24] which um traditionally the polyglots [82:26] team has used there. There's an awful uh [82:30] lot of well not a lot but there's a few [82:32] very incisive comments which come up but [82:35] um Matt is basically asking the question [82:37] what can we do with AI to make the whole [82:40] polyglots thing the translation um much [82:43] more straightforward and uh so if you're [82:45] into that go check it out right let's [82:48] have a little look um I've got one [82:50] product that I thought I'd mention this [82:52] week called review bird not come across [82:55] this before uh is for those of you that [82:58] are um using Woo Commerce stores and uh [83:01] I don't have a WooCommerce store and I [83:03] don't take product reviews so I don't [83:04] know how essential this kind of plug-in [83:06] would be but if you are and you want to [83:09] corral your product reviews go and check [83:11] out reviewbird.com [83:13] it allows you to manage all of that in [83:15] one place and given whenever I go online [83:17] and shop for things I do look at reviews [83:19] it's probably is quite an important part [83:21] of such a website right now that we're [83:24] um freaked ourselves out let's go to [83:26] this one to karma us all down, shall we? [83:28] This is Cloudflare. Their annual That's [83:32] not true. I think it's quarterly report. [83:34] Yeah, there you go. 2025 Q4 DDS threat [83:37] report. It's a It's a The title is [83:40] guaranteed to uh offer you sleep. But, [83:43] however, it says a record setting 31.4 [83:47] terabits per second attack caps a year [83:49] of impressive DDS DDoS assaults. [83:53] What didn't you know already from that? [83:55] But check it out. Like just some of [83:57] these numbers are absolutely ridiculous. [83:59] If you work for Cloudflare, God bless [84:01] you. I don't know how you guys manage [84:04] what you're doing, but uh here we go. Uh [84:06] the fourth quarter of 2025 was [84:08] characterized by an unprecedented [84:09] bombardment launched by the whatever the [84:12] heck that is. Arisuru Kim Wolf Botnet D. [84:16] >> They are famous botnet networks. [84:18] >> Oh, thank you. They have like uh they [84:20] have like over 200 last I checked 250 [84:24] million compromised devices handling. [84:27] >> Yeah. This was the thing that I was [84:29] saying earlier about the Google TV, you [84:31] know, the TV dongles and things like [84:32] that. Apparently, a lot of them are part [84:35] of this whole thing. Yeah. Um the [84:37] campaign targeted Cloudflare's customers [84:39] as well as Cloudflare as well as [84:40] Cloudflare's dashboard. [84:43] Now, I don't know if it's you, but I did [84:45] notice Cloudflare at the end of last [84:46] year going down a few times, like in in [84:49] a way that I'd never ever ever seen [84:51] Cloudflare go down before. And maybe [84:53] this explains it. Um, the hypervoluttric [84:57] HTTP DDoS attack succeeded. Just listen [85:01] to this number and try to imagine it for [85:03] a minute. Rates of 200 million requests [85:08] per [85:10] second. [85:11] I mean, what uh just weeks, that was [85:15] just weeks after a record-breaking 31.4 [85:18] terabits per second attack. Um and then, [85:22] you know, there's a bunch of charts [85:23] which ex which show fairly conclusively [85:26] that 2023 was lower than 2024. 2024 in [85:30] turn was significantly lower than 2025. [85:33] This graph looks logarithmic, by the [85:35] way, which means that by the year like [85:36] 2029, it's going to be infinite. [85:38] There'll be nothing else on the internet [85:40] apart from DDS attacks. It's just going [85:43] to go, "Oh, Tim." Yeah. Is that [85:45] plausible? [85:46] >> Um, yeah. I mean, [85:48] >> no. I I mean probably can answer this [85:52] much more clearly than I can, but yeah, [85:55] we we are we've already at the age where [85:57] we pretty much now [86:00] human traffic on the on the internet [86:04] that needs to be specified separate to [86:06] the web is such that we are now the [86:08] minority. Um and it will not be that [86:11] much longer before the web we are the [86:13] minority at which point yeah you you're [86:17] talking about automated systems talking [86:19] to automated systems. Um I I think these [86:22] numbers are are really interesting for [86:24] perhaps a completely different reason [86:25] and that's just how dominant Cloudflare [86:29] has become. [86:30] >> It's starting to worry me actually how [86:32] dominant Cloudflare is. [86:34] And as it happened, I think the times [86:36] that they were down, the big ones that [86:37] got in the news, each time was due to [86:40] conf literally uh conf errors rather [86:43] than actual I don't think they've gone [86:45] down from a compromise. But uh yeah they [86:49] are the just how dominant they've become [86:51] and how important they are to [86:52] infrastructure and uh whether that's a [86:55] good thing whether that's a bad thing uh [86:58] they are reaching that point where you [87:00] got to think about them and you you're [87:03] now in this position where you might as [87:06] you they've reached that tipping point [87:08] where up until recently there was like [87:10] well there's a little part way you might [87:12] you don't necessarily need to want to [87:14] use Cloudflare because you can use your [87:15] own but now If Cloudflare is down, so [87:19] much of the internet's down that so many [87:21] of the resources that you rely on for [87:23] your own website is probably down that [87:24] your website's probably down anyway, [87:27] >> right? Right. [87:28] >> So, you might as well use them. [87:31] >> The curious optimistic. [87:33] >> Yeah. But the the curious thing about [87:35] the Cloudflare thing is a lot of the [87:36] other platforms when they started to [87:38] gain dominance. I'll use Facebook as an [87:40] example [87:42] wrapped around that was a a growing [87:44] sense of like wow Facebook's like [87:47] there's a lot of wrong going on there. [87:50] There's a but I don't have that with [87:51] Cloudflare. I I don't get a sense of [87:53] there's a lot of wrong going on there. [87:55] It's just they're getting bigger and [87:57] bigger and bigger and so their failure [87:58] is just bad. I personally haven't read [88:01] enough. I've got no intuitions about [88:03] them being a, you know, and I'm doing [88:05] air quotes, a bad company doing [88:08] unethical things. Perhaps they are, but [88:10] I don't have that intuition. [88:12] >> They're a corporation at that scale. [88:14] There is always going to be some bad. [88:17] You can't be a corporation that size and [88:20] not [88:21] >> upset some person. [88:23] >> I have a solution, Tim. What we need to [88:24] do is put the is put Cloudflare [88:28] in a ring with Remus. it would lose [88:33] >> now that I think Remus can subscribe to [88:35] he's on team Cloudflare. I think maybe [88:38] they just shake hands and that would be [88:40] the end of it. Uh okay. Anyway, that is [88:42] really sanguin stuff. You know, the [88:44] numbers and the sort of the future that [88:46] we're painting ourselves into the corner [88:48] of. Kind of interesting, right? Some [88:50] more security stuff very very quickly. [88:52] CSS got hacked. [88:55] How is this possible? Go and read the [88:57] article. It's on Medium. It's kind of [88:58] interesting. It's complicated and [88:59] interesting and it's the way that the [89:01] Chrome browser uh pares. It's not C, you [89:05] know, you don't have to actually worry [89:06] about writing CSS apart from in some [89:09] really curious cases. But uh I've never [89:11] seen a vulnerability involving CSS [89:13] before. [89:14] >> It's just as well we have lots of [89:16] different browsers and we're not all [89:17] based on Chromium based systems, isn't [89:19] it? [89:19] >> Well, I know you are still So you still [89:23] using Firefox over there? Uh yes, except [89:26] for when doing broadcast to you and my [89:29] >> That's right. Because these platforms [89:30] won't allow it. Yeah, there's you and [89:32] your you and your seven other mates [89:34] still using I I love I love I really [89:38] want Mozilla to succeed. I listened to a [89:40] podcast episode with the the now new CEO [89:44] of Mozilla and he was trying to paint a [89:47] positive spin on their new AI [89:48] initiatives and I've got to say it [89:50] didn't wash too well with me. I I got [89:53] the bad feels from it. It was like sort [89:56] of like greenwashing what is a necessary [89:59] but horrible awkward evil that they're [90:02] having to deal with. Yeah. You in [90:04] agreement, Tim, a little bit. [90:05] >> Oh, yeah. Absolutely. [90:06] >> Okay. Great. Yeah. [90:08] >> Um [90:09] >> I think they released new feature to [90:11] disable all the the AI features. [90:15] >> Yes. Well, at least that's a thing, [90:16] right? You can go in there and switch [90:18] them all off, but ne necessarily they've [90:20] they're going to be enabling them by [90:22] default. But also, I just want a browser [90:25] clear of all that really. I want the [90:26] browser to just browse the internet [90:28] based upon where I click and nothing [90:30] more. [90:30] >> Obviously, there are the nice thing is [90:32] that there are just like there are [90:33] different Chromium browsers. There are [90:35] different Firefox variants out there and [90:37] people are working on [90:39] >> a good one. Libra Wolf, [90:41] that's very good one. If anyone who's [90:44] interested in privacy, Libra, what was [90:46] that one called? I [90:47] >> think called Libra Wolf. [90:49] >> Oh, okay. [90:50] >> As in like Libra Office, Libra free. [90:53] >> Yeah. Yeah. Perhaps from the Yeah. Libra [90:55] Wolf. It's a Libra Wolf variation. [90:59] >> Yep. I downloaded one not that long ago [91:02] which I didn't actually use because [91:03] there were significant things in the UI [91:05] which I didn't really like. But it was [91:06] curious because it was a it was a [91:08] Firefox clone but it was also trying to [91:11] enable Chrome and the Firefox apps uh uh [91:16] extension ecosystem. [91:18] >> Was that Waterfox? [91:20] >> It was I don't think it was. Um [91:22] >> there certainly the no AI pushing [91:25] heavily one. [91:26] >> Okay. But it was just the idea that [91:28] there are a few Chrome extensions that [91:30] I'm really familiar with and I would [91:32] love to be able to use and there's no [91:33] equivalent Firefox one. Um, but this [91:36] browser I thought might be able to do [91:37] both but it didn't manage it. There was [91:39] >> Firefox will install nearly all Chrome [91:42] extensions. [91:43] >> Okay. [91:44] >> Just as a complete aside if you because [91:46] they they they do share a commonality [91:48] for their APIs. [91:49] >> Okay. Okay. Uh so some comments coming [91:52] in about that. uh the entire Firefox [91:54] community. Oh, that's left. He's there. [91:57] Look, there there is the Firefox. [91:59] There's the UK Firefox community. [92:01] >> And do not worry, I am indeed up in [92:03] arms. [92:04] >> Yeah, is up in arms about that. I used [92:06] to be on their Sumo team. There's no [92:08] connection between the CEO and those [92:11] marketing the product. Okay. Um Right. [92:14] Okay. I think very very quickly we'll [92:16] just go through these. Uh I know time is [92:18] short and I don't want to hold the the [92:20] guests uh to stay for any length of time [92:23] longer than we said we would. Um if you [92:26] are using manage WP owned by GoDaddy, [92:29] they have now forged a relationship with [92:31] Patchack. I'll just mention that you can [92:33] go and read about that yourself. The I [92:36] don't know if this is the final round, [92:37] but if you're going to go to Word Camp [92:39] Asia or have intuitions that you might [92:41] like to go to Word Camp Asia, the fifth [92:42] round of speakers has been announced. [92:45] And then [92:47] uh we had the dancing robots already. D [92:50] I'm just going to drop this one and say [92:53] make of this what you will. This like [92:56] just look at that for a minute and pause [92:58] it. Just take in what you're seeing [93:02] and then tell me that like if you needed [93:05] proof of the existence of God or [93:06] something like that, this might be it. [93:09] Uh this is a fossil. like nature made [93:12] that and it was found by somebody on a [93:14] beach in and believe it or not they [93:16] found it on the island called Holy Holy [93:20] Island uh in Lindesvan and that's so [93:23] obviously somebody's mouth the human did [93:26] that there's no doubt about it or the [93:28] aliens the aliens left it there that's [93:30] probably what happened cuz look it's got [93:32] like the gums the pink gum bits and then [93:34] the teeth and then the Anyway apparently [93:37] not apparently it's a fossil it's on the [93:39] BBC website. Smiling fossil discovered [93:42] on Holy Island [93:44] drawn by teenager two weeks before with [93:47] crayon. [93:49] It doesn't actually say that, but I like [93:51] the idea. Okay, there we go. That's it. [93:53] We've run out of time. We've actually [93:54] overshot the amount of time that we [93:56] should be doing this show. Thank you uh [93:58] to you for joining in the comments. Uh [94:00] we await the obvious wrestling match [94:03] between Matt and Remus. That'll be [94:05] interesting to see. Maybe bring that [94:06] >> battle. [94:07] >> The rap battle. That's right. Yeah. No [94:10] violence. [94:11] >> No violence. The no violence rap mod [94:14] wrestling match. That's what I heard. [94:17] >> How about a break dance competition [94:19] instead? [94:20] >> Oh, what? [94:21] >> A break dance. A break dance [94:23] competition. [94:23] >> Yeah. Okay. Yeah. To to be honest with [94:26] you, what s what would satisfy me right [94:28] at this moment because I've been [94:29] watching so much of it is if they both [94:31] just had a go at curling. just do the [94:35] bit of curling and see who's the uh [94:37] >> who's the [94:38] >> I like the idea of the I feel I feel [94:40] feel that that still very much uh leans [94:42] more toward Remicus than it does Matt. [94:44] >> Yeah. Yeah. Okay. We keep present. Yeah. [94:46] Okay. Well, how about the my surname is [94:49] Cromwell. [94:50] >> Matt could hire that Australian uh one [94:53] person who competed in the last [94:54] Olympics. You remember her? [94:56] >> He could hire her to to Yeah. [95:00] watch [95:02] >> to do the curling cuz that would be the [95:05] curling break [95:07] >> can do all the moves and then finally [95:09] release it. I accidentally caught sight [95:12] of that Australian lady doing the [95:14] whatever that event is. Is it like [95:16] street dance or something like that? [95:17] >> It's broken dancing is what it was break [95:20] dancing. [95:20] >> It was on there was obviously a news [95:22] article. It was on in a shop that I was [95:24] in the other day and I watched it and it [95:25] really even though I know nothing about [95:28] that event, I have no chops about what's [95:30] credible and what's not. I was watching [95:32] it like, oh, it's so weird. It's so so [95:36] bad. Anyway, there we go. Thanks to our [95:39] team. Patricia says, appreciate it. [95:42] We'll be back this time next week, I [95:44] hope. Um, should be back this time next [95:48] week, but uh thank you. I'm just going [95:49] to go around the houses and say thank [95:50] you to Tim Nash. [95:52] >> There he is. Thank you to Mark [95:54] Benzakane. It's still like crazy o'lock [95:57] in the morning over there. Appreciate [95:58] that. [95:59] >> And also to Onar. Thank you so much for [96:02] joining us. Hopefully Onar, you'll come [96:03] back and join us on a less weird [96:06] episode. [96:06] >> They're always weird. They're always [96:08] weird. [96:09] >> They're not as weird. [96:10] >> Don't Don't fall for it. Every single [96:12] episode is weird. [96:13] >> Okay. All right. Okay. I probably should [96:15] rate that. [96:16] >> I I love it, though. This what makes [96:17] this fun, right? It's weird. So, [96:19] absolutely. [96:20] >> Yeah, I agree. [96:21] >> And uh we'll be back this time next [96:23] week. So, if the three of you want to [96:25] stick around and have a little bit of a [96:26] natural after this, I would welcome [96:28] that. But uh we will see you next week. [96:30] Thanks for joining us in the comments. [96:31] Appreciate it. Bye-bye. Bye-bye. ## Highlights - talking about WordPress. He was talking about sort of agency life and things like that. But the one of the things that caught my attention was his position on AI. Now Andy is very principled about AI and he's very very forthright. He doesn't want anything to do with it. He doesn't want anything he doesn't want his agency touching it. So the content that they produce, the code that they write, the websites that they launch is all humanmade. As soon as those words came out of my mouth, I realized the error. Uh it was ([View Highlight](https://read.readwise.io/read/01kjbc3tb58m165yngrv4pjdpq)) - Tags: 👻 ai highlighted - anyway, the knock-on effect was as soon as he started saying the quiet thing out loud, I refuse to work with AI, suddenly his roster got filled up again. So it turned out that for a small subset of people and he presumably is an early mover that that rounded off his 2025 really well. Now we'll have to see how long that continues, but it turned out that that principle resonated with a lot of people who needed a website. Um but we'll have to wait and see how that that ([View Highlight](https://read.readwise.io/read/01kjbc3tb8x8bnm4yfg8v1t7tr)) - Tags: 👻 ai highlighted - that way as well. So >> thank you. Okay, that's really interesting. Uh I didn't say the title of the piece and I have I'll read it in because if anybody's uh listening to this primarily this is on the automatic.com. So with 2T's blog and it's called a modest server centric development proposal. So you can go and look for that. It was written on the 17th of Feb this year. So go and check that out. Right. And there was another one coming from you. Sorry I apologize. I didn't let Mark and uh Onar have their ([View Highlight](https://read.readwise.io/read/01kjbc3tba4y3bn32fx8zw517p)) - Tags: 👻 ai highlighted - you every few days it seems you drop a little nugget of something that you've played about within WordPress 7 and uh, yeah, I'm very grateful for you doing that. Appreciate it. So anyway, go and check that out. Just a quick mention, uh, friend of this podcast, he's been on it lots and lots of times, no doubt will be in the future as well, Remus Dice. Um, he's launching a new service. He's he's had a service for a little while. uh truer than north. But I I won't try to pause this. It's a fairly lengthy written article, but he is um sort of ([View Highlight](https://read.readwise.io/read/01kjbc3tbe3dh0sbpy6vr1grxs)) - Tags: 👻 ai highlighted - I know Courtney's joining us in the comments. She probably has a lot more uh sense to say about this because she is one of the founders as well. Um and yeah, the endeavor of this project is to is to you know sort of grow the community in a fair, equitable, transparent way. And uh I feel I've had a few chats with uh Chris Reynolds. I don't know that I've ever spoken to Jonno Alderson, but my the only experience that I've had with Jonno was watching him speak um at a WordPress ([View Highlight](https://read.readwise.io/read/01kjbc3tbhhhr9x741x7xahn5v)) - Tags: 👻 ai highlighted - be joining the board and helping to push WordPress towards a healthier, more equitable future." This is on the repository uh blog, so you can go and check it out over there. Uh anything from you three, excuse me, about that? >> Not about that, but can we just take a moment? So, uh, you can't see this on the podcast, but the amazing picture that the repository found of Chris, which is like you can't even see his face. He's just >> Which one is it? Is this him here? ([View Highlight](https://read.readwise.io/read/01kjbc3tbm5yvv9fhrgvgdg6k5)) - Tags: 👻 ai highlighted - I'm surprised he didn't put the word fight in quotes while he was at it. >> Yeah, that's right. Okay, let's see if Matt Cromwell drops in and and offers it offers it off. Okay, right. Let's get into some sort of security stuff. There's a few bits and pieces here. Um, it's totally out of my pay grade this, but I did think this was kind of interesting. So, this is talking about AI and security all in one go. So, somebody that I followed on social media for a little while, this is on Masttodon, uh Kevin Bowmont, he says, ([View Highlight](https://read.readwise.io/read/01kjbc3tbq5es1bv230f4dy5sk)) - Tags: 👻 ai highlighted - that was really interesting and thanks Mark for re rediscovering it. >> Ju just as an aside, I've just put a a link in the guest chat uh which probably you've already linked to in past episodes. Uh but if we're going to have if we're going to have AI fight, so at least we've now got the the option. >> Okay, let me let me try. Is this Can I put this on the screen? Is it safe? >> Yeah, it's safe. >> Right, hold on. Okay, we're all all having a giggle, are we? Okay, here it comes. Oh, no. I haven't fe This is insane. I ([View Highlight](https://read.readwise.io/read/01kjbc3tbt3dj1jadj1an65cy9)) - Tags: 👻 ai highlighted - couple of pieces which I'm not going to have time to get to. Uh so let's see where we get to now. Oh, this is another piece that came from Matt. Um if you're in the polyglots um ecosystem in the WordPress space, this might be worth looking at. It is AI related again as everything seems to be and it's about using uh AI to achieve the the things which um traditionally the polyglots team has used there. There's an awful uh ([View Highlight](https://read.readwise.io/read/01kjbc3tbxbxxk70en1ptqxgr5)) - Tags: 👻 ai highlighted - >> now that I think Remus can subscribe to he's on team Cloudflare. I think maybe they just shake hands and that would be the end of it. Uh okay. Anyway, that is really sanguin stuff. You know, the numbers and the sort of the future that we're painting ourselves into the corner of. Kind of interesting, right? Some more security stuff very very quickly. CSS got hacked. How is this possible? Go and read the article. It's on Medium. It's kind of interesting. It's complicated and interesting and it's the way that the ([View Highlight](https://read.readwise.io/read/01kjbc3tc0x2s607dd0rm6qx6y)) - Tags: 👻 ai highlighted ## 🌱 Spawned Notes - [[Lit · WordPress collaborative editing transforms content workflows]] - [[Lit · Open source AI models closing quality gap rapidly]] - [[Zet · Server-side development reduces JavaScript complexity]] - [[Zet · Password manager adoption essential for security]] - [[Flt · Test WordPress 70 beta collaborative features staging environment setup]] - [[Flt · Research Kitten TTS integration for accessibility 25MB model evaluation]] - [[Flt · Review MainWP anniversary insights 12-year sustainability factors]]