Open Source Might Change Forever - CourtneyR.dev's 2nd Brain

Open Source Might Change Forever ![rw-book-cover](https://img.youtube.com/vi/2bYPsQvdfe0/maxresdefault.jpg) 🔗 **Source:** [Open Source Might Change Forever](https://www.youtube.com/watch?v=2bYPsQvdfe0) > [!note]+ Document Notes > tags: [open-source, ai-development, web-frameworks, test-driven-development] > author: ThePrimeTime > source: YouTube > reference: https://www.youtube.com/watch?v=2bYPsQvdfe0 > date created: 2024-12-19 10:30 > > # 🏷️ Title: Open Source might change forever >  ## 🧰 Action Inbox - [H1] Consider implications of AI-driven framework cloning for open source sustainability #open-source - [H2] Evaluate whether to keep test suites private to prevent AI copying #test-driven-development - [H3] Research Cloudflare's Vinxi performance claims vs Next.js #web-frameworks ## 🧭 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 - [Cloudflare Vinxi Project](https://vinxi.dev) - [Next.js Documentation](https://nextjs.org) - [SQLite Test Harness](https://sqlite.org) ## 📝 Summary - Cloudflare used AI to clone Next.js in one week using 1,700 Vitest tests and 380 Playwright tests, creating Vinxi - Claims 57% smaller bundles and 4x faster builds, though with reduced feature set (no static pre-rendering) - Raises fundamental questions about open source sustainability when AI can rapidly clone projects using public tests - SQLite's private test suite model may become necessary for protecting open source competitive advantages - Government sites already running experimental Vinxi in production raises deployment concerns ## 📖 Main Content Following Anthropic's AI-generated C compiler demo, Cloudflare demonstrated a more practical application by rebuilding Next.js using AI agents. Unlike Anthropic's marketing-focused example, Cloudflare's Vinxi project represents a functional framework clone that's already deployed in production environments. The technical achievement involves processing Next.js's public test suite to recreate the framework's functionality with different architectural choices. However, the performance claims require scrutiny - the 57% smaller bundles and 4x faster builds come with significant caveats around missing features like static pre-rendering. This development pattern suggests a potential shift in open source strategy, where maintaining competitive advantages may require keeping critical test suites private, similar to SQLite's TH3 test harness approach. ## 🗒️ Notes #### ⏱️ [00:00](https://youtu.be/2bYPsQvdfe0?t=0) — [H1] Anthropic's C compiler marketing vs reality - Key idea: Anthropic claimed building C compiler "from scratch" but used existing test suites - Evidence: "Hey, it wasn't really from scratch. You used a bunch of test suites and everything" #### ⏱️ [01:15](https://youtu.be/2bYPsQvdfe0?t=75) — [H1] Cloudflare's Next.js clone project - Key idea: Cloudflare rebuilt Next.js using AI in one week with dramatic performance claims - Evidence: "57% smaller client bundles, 4x faster builds" #### ⏱️ [02:30](https://youtu.be/2bYPsQvdfe0?t=150) — [H2] Next.js hosting complexity background - Key idea: Next.js is notoriously difficult to host outside of Vercel - Evidence: "Open Next...was created with the sole purpose of making hosting Next.js not on Vercel easier" #### ⏱️ [03:45](https://youtu.be/2bYPsQvdfe0?t=225) — [H1] AI methodology using public tests - Key idea: Used 1,700 Vitest tests and 380 Playwright tests to recreate Next.js functionality - Evidence: "Cloudflare used these 1,700 Vitest tests and 380 Playwright end-to-end tests to effectively recreate Next.js" #### ⏱️ [04:30](https://youtu.be/2bYPsQvdfe0?t=270) — [H3] Government production deployment concerns - Key idea: Experimental framework already running on government beta sites - Evidence: "They're already running Vinxi in production...let's not use experimental frameworks on the gov" #### ⏱️ [05:15](https://youtu.be/2bYPsQvdfe0?t=315) — [H2] Test coverage reality check - Key idea: Only covers 13% of Next.js test cases despite 94% API surface claim - Evidence: "13% dev, 20% end-to-end, and 10% production. There are 13,708 test cases" #### ⏱️ [06:30](https://youtu.be/2bYPsQvdfe0?t=390) — [H2] Missing static pre-rendering feature - Key idea: Vinxi lacks key Next.js feature of build-time static rendering - Evidence: "Vinxi does not support static pre-rendering at build time" #### ⏱️ [07:45](https://youtu.be/2bYPsQvdfe0?t=465) — [H1] Traffic-aware pre-rendering innovation - Key idea: Cloudflare's approach builds only frequently visited pages using traffic data - Evidence: "12,000 unique paths, 184 pages covered 90% of the traffic" #### ⏱️ [09:00](https://youtu.be/2bYPsQvdfe0?t=540) — [H2] Skepticism about bundle size claims - Key idea: 57% smaller bundles seems implausible without missing features - Evidence: "I don't believe you...You can't just have 50% smaller and everything else the same" #### ⏱️ [11:30](https://youtu.be/2bYPsQvdfe0?t=690) — [H1] Open source vulnerability implications - Key idea: Public tests enable AI-driven competitive cloning of open source projects - Evidence: "They were effectively able to brain drain it like Silicon Valley and boom, they have their own version" #### ⏱️ [12:45](https://youtu.be/2bYPsQvdfe0?t=765) — [H1] SQLite's private test strategy - Key idea: SQLite keeps critical test suite TH3 private to prevent copying - Evidence: "TH3 sources...differ from the others in being proprietary. The TH3 code is only accessible to SQLite developers" #### ⏱️ [14:00](https://youtu.be/2bYPsQvdfe0?t=840) — [H2] Future of company-backed open source - Key idea: Companies may need to keep tests private to maintain competitive advantages - Evidence: "Do you think we're going to see the day where...company-backed open source no longer...makes public any sort of testing?" #### ⏱️ [15:30](https://youtu.be/2bYPsQvdfe0?t=930) — [H3] Maintainability concerns for AI-generated code - Key idea: Long-term maintenance of AI-cloned frameworks remains questionable - Evidence: "I wonder how maintainable is it? Are they permanently just required to live in the slop universe?" ## 📚 References - [Cloudflare Vinxi Framework](https://vinxi.dev) - [Next.js Official Documentation](https://nextjs.org/docs) - [SQLite Test Harness Documentation](https://sqlite.org/th3.html) - [Open Next Project](https://open-next.js.org) ## 🗣️ Paraphrase - Cloudflare used AI to recreate Next.js in a week by feeding it the framework's public test suite, claiming significant performance improvements but missing key features like static pre-rendering - The ability to rapidly clone open source projects using AI raises questions about whether companies should keep their test suites private to maintain competitive advantages, similar to SQLite's approach - While the technical achievement is impressive, concerns remain about the maintainability of AI-generated code and the broader implications for open source sustainability ## 📰 Newsletter **AI Cloning Frameworks: What It Means for WordPress Development** Cloudflare just demonstrated something that should make every open source maintainer pause: they rebuilt Next.js in one week using AI and the framework's public test suite. While their performance claims are impressive (57% smaller bundles, 4x faster builds), the real story is what this means for the future of open source development. For WordPress developers, this raises important questions about our ecosystem's sustainability. If AI can rapidly clone complex frameworks using public tests, should projects consider protecting their test suites? WordPress Core's extensive test coverage is a strength, but it could also become a vulnerability in an AI-driven development landscape. The WordPress community's collaborative approach has always been our superpower, but we may need to evolve our strategies while preserving our open source values. Resources: [WordPress Core Tests](https://make.wordpress.org/core/handbook/testing/), [WP-CLI Testing](https://wp-cli.org/), [WordPress Coding Standards](https://developer.wordpress.org/coding-standards/) ## ✅ Action Items #### Community - Maintainers: Evaluate test suite exposure risks, develop AI-aware contribution policies - Contributors: Understand implications of public test contributions in AI era - Ecosystem: Coordinate cross-project response to AI cloning concerns #### WordPress - Core: Assess test suite vulnerability, establish AI development guidelines - Core Performance: Monitor AI-generated optimization claims skeptically - Community: Develop educational content about AI's impact on open source - Documentation: Create guidance for AI-resistant development practices - Meta: Establish policies around AI-generated code contributions #### OSPO - Policy & Governance: Draft AI cloning impact assessments for contribution policies - License Compliance: Review licensing implications of AI-trained reproductions - Security Management: Evaluate test suite exposure as security consideration - Community Engagement: Facilitate discussions on AI's open source impact #### FAIR - Technical independence: Ensure AI cloning doesn't compromise project autonomy - Policies: Develop transparency requirements for AI-assisted development ## 🧪 Hypotheses - Open source projects will increasingly adopt private test suites to prevent AI cloning - AI-generated framework clones will initially show performance gains but struggle with long-term maintainability - The competitive advantage of open source companies will shift from code to community and expertise - Government and enterprise adoption of AI-cloned frameworks will create new security and compliance concerns ### 📊 Methodology - Comparative analysis of Vinxi vs Next.js feature parity over time - Survey of open source maintainers on test suite privacy strategies - Performance benchmarking of AI-cloned vs original frameworks - Legal analysis of AI-generated code licensing implications ### ❓ Questions - How will WordPress Core balance test transparency with protection from AI cloning? - What percentage of Next.js functionality can AI recreate with current test coverage? - Will the open source definition need updating to address AI-generated reproductions? ### 🔮 Future Research - Impact of AI cloning on open source funding models and sustainability - Development of AI-resistant open source licensing frameworks - Community governance models for AI-assisted development ### 💡 Implications - Practice: Developers may need to evaluate AI-cloned alternatives against original projects - Policy: Open source licenses may require updates to address AI training and reproduction - Security: AI-generated code introduces new audit and vulnerability assessment challenges - Education: Developer training must include AI-assisted development ethics and practices - Funding: Open source monetization strategies must evolve beyond code differentiation ### 🧠 Learning Styles Assessment Reading level estimate: 10th Grade (Flesch-Kincaid) #### 👀 Visual learners - Present: Performance comparison charts, test coverage diagrams - Improve: Architecture diagrams showing Vinxi vs Next.js differences, timeline of AI development #### 🔊 Aural learners - Present: Video format with clear narration and examples - Improve: Podcast-style discussion of open source implications #### 🗣️ Verbal learners - Present: Clear analogies ("brain drain like Silicon Valley"), defined technical terms - Improve: More framework comparisons, expanded metaphors for AI cloning #### 🧩 Kinesthetic learners - Present: Real production examples, actionable concerns - Improve: Hands-on comparison labs, interactive framework demos #### 🧠 Logical learners - Present: Performance metrics, test coverage percentages, logical progression - Improve: Decision trees for test suite privacy, cost-benefit analysis #### 🧑‍🤝‍🧑 Social learners - Present: Community impact discussion, collaborative implications - Improve: Discussion guides for developer teams, community response frameworks #### 🧘‍♀️ Solitary learners - Present: Technical deep-dives, individual analysis opportunities - Improve: Self-assessment tools for AI impact, personal project evaluation checklists #### 🌿 Naturalistic learners - Present: Ecosystem impact thinking, natural evolution parallels - Improve: Open source ecosystem health metrics, sustainability frameworks ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Clear heading structure, descriptive section labels - Improve: More detailed alt text for performance comparisons, ARIA labels for complex concepts #### Keyboard navigation - Present: Logical content flow, clear section breaks - Improve: Skip links for technical sections, better focus management #### Contrast and readability - Present: High contrast text, reasonable line length - Improve: Ensure code examples meet contrast requirements #### Text resizing and reflow - Present: Responsive design accommodating zoom - Improve: Test technical content at 200% zoom #### Media and captions - Present: Video includes transcript content - Improve: Detailed captions for technical terms, audio descriptions #### Forms and interactions - Present: Simple navigation structure - Improve: Accessible feedback forms for community input ### 🌍 DEIB Assessment #### Cultural breadth - Present: Global open source perspective, international impact consideration - Improve: Examples from diverse development communities, non-US framework examples #### Inclusive language - Present: Technical terms explained, avoiding unnecessary jargon - Improve: Plain language summaries, cultural context for open source concepts #### Neurodiversity support - Present: Clear structure, chunked technical information - Improve: Visual breaks in complex sections, summary boxes #### Representation - Present: Technical focus with broad community implications - Improve: Diverse developer perspectives, varied use cases #### Psychological safety - Present: Balanced technical criticism, respectful tone - Improve: Acknowledgment of different skill levels, welcoming language ### ✅ Logical Validity #### Strong reasoning - Inductive strength: Pattern recognition from Anthropic to Cloudflare examples shows emerging trend - Causal reasoning: Clear connection between public tests and AI cloning capability - Analogical strength: SQLite private test comparison provides relevant precedent #### Fallacies - Hasty generalization: Extrapolating from two examples to entire open source future - Appeal to incredulity: Dismissing 57% performance claims without thorough analysis - Slippery slope: Assuming all open source will move to private tests #### Tone of voice - Overall tone: Skeptical but engaged, technically informed with casual delivery, appropriately concerned about implications ### 🛡️ DARVO Analysis - Deny: No denial patterns present - Attack: No attacks on accusers - Reverse victim/offender: No victim/offender reversals - Emotional cues and rhetoric: Healthy skepticism and humor rather than manipulative rhetoric ## 🎯 Conclusion The rapid AI-driven cloning of Next.js by Cloudflare represents a watershed moment for open source development. While the technical achievement is impressive, it forces fundamental questions about how open source projects can maintain competitive advantages and sustainable funding models in an AI-enabled world. For the WordPress ecosystem, this development suggests we need proactive strategies that balance our commitment to openness with practical sustainability concerns. The future may require more nuanced approaches --- ## Full Transcript [00:00] Last week, Anthropic released this [00:01] marketing video in which they're like, [00:02] "Hey, we built a C compiler from [00:04] scratch." And you know, like we just [00:06] walked away and 2 weeks later it was [00:07] like poof, C compiler and everything [00:09] worked. I could build Linux. It was easy [00:12] peasy. Engineering's over. Now, of [00:14] course, this led to a bunch of people to [00:16] kind of actually say, "Well, hey, it [00:17] wasn't really from scratch. You used a [00:18] bunch of test suites and everything." [00:20] But Cloudflare, when they saw that, they [00:22] said, "Well, wait wait wait a second. [00:24] Hold my beer. I got an idea." And they [00:26] decided they're going to rebuild [00:28] Next.js, which is, if you don't about [00:30] Next.js, just imagine a framework and [00:33] you put a framework in a framework. [00:34] That's effectively what Next.js is. They [00:37] call it a meta framework, okay? Very [00:39] metaphysical, okay? I actually I don't [00:41] even get that joke. And they make [00:43] outrageous claims about it. 57% smaller [00:46] client bundles, 4x faster builds. All [00:49] right, so we're going to obviously have [00:51] to go over this. We're going to talk [00:52] about these claims. We're going to [00:54] actually kind of understand the whole [00:56] purpose to everything. And even more so, [00:58] we should probably talk about the [01:00] implications of this. But first, I don't [01:02] know if you know this, on March 27th, [01:04] I'm going to be in San Francisco. I'm [01:05] going to be doing an event for Century. [01:06] If you want to come, sign up. The link [01:08] will be in the description. It's going [01:09] to be a lot of fun. A lot of founders [01:10] and engineers are all going to be [01:12] hanging out at the Chase Center. [01:13] Sportsball, am I right, boys? [01:16] Yeah? You know me and that sportsball. [01:18] Hey, but if we're talking pickleball, [01:20] dog, I'll see you on the court. [01:22] Anyways, link will be in the [01:23] description. Come hang out. It's going [01:25] to be awesome. So, first to kind of [01:26] understand what's happening, if you [01:27] don't know anything about Next.js, it's [01:29] chronically been known on the internet [01:31] as a very difficult thing to host if [01:32] you're not on Vercel. Open Next right [01:34] here, which was created, of course, by [01:36] Dax. Look at Dax. Look at all that money [01:38] on him. Classic Dax right there. It was [01:39] created with the sole purpose of making [01:41] hosting Next.js not on Vercel easier. [01:44] And there's a lot of reasons why this is [01:46] difficult. No need to go into them. Just [01:48] know that there's some hurdles one has [01:49] to jump over to make this easy. Now, [01:51] what makes this really interesting is, [01:52] if you remember with Anthropic, [01:55] they used the 37 years of engineering [01:57] test to be able to produce a C compiler [02:00] that they probably already had memorized [02:01] in the weights. But this one is is [02:03] different because Cloudflare used these [02:05] 1,700 Vitest tests and 380 Playwright [02:08] end-to-end tests to effectively recreate [02:11] Next.js, but instead of having kind of [02:13] the proprietary or the Turbopack or [02:15] whatever you would call that kind of [02:17] build style, instead it was something [02:18] that'd be a lot more friendly to say [02:20] Netlify or Cloudflare. And in 1 week, [02:22] they were able to actually like get it [02:24] to work. Now, this part actually right [02:26] here terrified me. We've been working [02:28] with the National Design Studio team [02:29] aiming at modernizing every government [02:31] interface on their beta sites, cio.gov. [02:34] They're already running Vinxi in [02:35] production. Okay, hey, brothers. [02:37] I mean, hey, Cloudflare, I love you, but [02:39] >> [laughter] [02:40] >> let's not use experimental frameworks on [02:42] the gov. [02:43] Oh, my gosh, what are we doing? Dude, [02:45] it's a fun time. Hey, we all get to [02:47] laugh during the slopocalypse, okay? [02:50] Actually, I have to interrupt myself [02:51] quickly. A couple hours later, I was [02:52] doing some more reading on this and I [02:54] realized that I kind of misrepresented [02:56] stuff right here. 94% of the Next.js 16 [02:59] API surface sounds very intense, but if [03:02] you go and look at how much of the [03:03] actual Next.js test coverage this [03:05] Cloudflare port actually has, it's 13% [03:08] dev, 20% end-to-end, and 10% production. [03:11] There are 13,708 [03:14] test cases. And there's only 1,700 and [03:16] 380 end-to-end ones. So, very [03:19] interesting. 1 week of having an agent [03:21] go hog wild. I mean, it's impressive [03:23] it's made it this far and people are [03:24] actually using it. Still a little [03:25] worried about that government thing, [03:27] though. Okay, I just just throwing it [03:29] out there. Okay, so let's talk about the [03:30] claims. 4x faster builds and 57% [03:33] smaller. So, we'll get to the second one [03:35] in just a moment. The first one, I can [03:38] understand why this happens. First off, [03:40] they have just different build systems, [03:41] so maybe Rollup just truly is that much [03:43] more performant than say Turbopack. I'm [03:45] not saying which one's actually better [03:47] here, but let's just pretend like that [03:49] is a real thing. But there's a lot of [03:50] things that it doesn't do. So, this is a [03:52] very important part. This Vinxi or [03:54] whatever they're calling it, this this [03:56] new Next.js Cloudflare edition, it [03:58] doesn't do quite everything that Next.js [04:01] does. Right here, Vinxi does not support [04:03] static pre-rendering at build time. In [04:05] Next.js, pages without dynamic data get [04:07] rendered during next build and served as [04:09] static HTML. If you have dynamic routes, [04:12] you use generate static params to [04:13] enumerate which pages to build ahead of [04:15] time. Vinxi doesn't do that yet. It's [04:18] obviously doing less work, but that's [04:21] not quite the whole story. This I [04:22] actually find this part to be super [04:24] interesting. Introducing traffic-aware [04:26] pre-rendering. Next.js pre-renders every [04:28] page listed in generate static params [04:30] during the build. A site with 10,000 [04:33] product pages means 10,000 renders at [04:35] build time, even though 99% of those [04:36] pages may never receive a request. Build [04:38] scales linearly with page count. This is [04:41] why Next.js sites end up taking [04:42] 30-minute builds. Okay, I've [04:44] ever been there. I didn't know about [04:46] this. I have been in places where there [04:48] are 45-minute TypeScript builds. I do [04:50] get this. So, we take traffic-aware [04:52] pre-rendering, TPR. By the way, [04:54] front-end people love abbreviations. We [04:58] already I mean, they just got done [04:59] saying ISR just just a little bit ago, [05:01] okay? They're all about these things. [05:03] So, we built traffic-aware [05:04] pre-rendering. It's experimental today [05:06] and we plan to make it the default once [05:08] we have a more real-world testing. The [05:10] idea is simple. Cloudflare already is a [05:11] reverse proxy for your site. They [05:13] already know everything that's being [05:14] visited. So, why not use that [05:15] information and just build the things [05:18] you need. And as you can see right here, [05:20] 12,000 unique paths, 184 pages covered [05:23] 90% of the traffic. Actually pretty [05:25] cool. Pretty based, I'd say. All right, [05:27] so the next two parts I'm actually very [05:28] excited to talk about. So, the first one [05:29] being the 57% smaller. Okay, [05:33] I'm just going to throw this out there. [05:34] I don't believe you. And hold on, just [05:36] let me just let me just let me just back [05:38] that up for a second, okay? I've been [05:39] around the block for a while and I've [05:41] seen a lot of performance claims. And I [05:43] have been a part of grand rewrites. And [05:46] I've seen things be super fantastic and [05:48] then become, well, equal if not worse [05:50] than the other one. So, there's only a [05:52] couple possibilities that we're seeing. [05:55] First, we are seeing the early days of [05:57] Vinxi, meaning that not all the features [06:00] are there, not all the things that made [06:02] Next.js big are there. So, as time goes [06:05] on, size will go up and then they'll [06:07] become much closer, like within 5% of [06:09] each other. I think this is probably the [06:10] most likely answer is that Vinxi is just [06:13] not doing things and that saves them a [06:16] lot. Because the only other answer is [06:18] that Next.js has so thoroughly screwed [06:22] up, they're sending effectively 2x more [06:25] than they need to. And honestly, I [06:28] refuse to believe that. Like just ain't [06:31] no way. I did corner a Cloudflare [06:33] engineer and I did ask them many a [06:36] times. And at the end of the [06:38] conversation, he's like, "It's real." [06:39] And I was like, "I don't believe you." [06:41] That was my only answer. I just No. You [06:43] can't just do You can't just have 50% [06:46] smaller and everything else the same. I [06:48] just [06:49] not believing it. But in the small [06:51] chance that it's actually true, that's [06:53] incredible. What did Next.js do to screw [06:57] this up? [06:58] But the performance does seem to be like [07:01] something's going on because if you look [07:02] right here, this slop fork is faster and [07:05] smaller. You can see right here, [07:07] people's P99 times are actually [07:10] dropping. Okay, so [07:12] there's a there there. I have no idea of [07:14] the validity of the data or anything, [07:16] but if this is to be believed by this [07:18] Cloudflare employee, which again, I [07:20] mean, that's trusting, you know, the old [07:22] fox with the hens, as they say. But if [07:24] this is to be true, then then oh my [07:26] gosh, that's actually a dramatic [07:28] improvement. But the most important [07:30] topic is what are the implications from [07:32] this? Because the C compiler was just [07:34] like, oh, we all kind of laughed at [07:35] Anthropic because of how much they just [07:37] misrepresented it. Whereas the [07:39] Cloudflare one, they're like, "No, dog, [07:41] we're running this in production." Yeah, [07:42] people are just using it. Okay, so this [07:44] one's a bit more real. And they did the [07:46] exact same strategy. The tests were [07:49] open. They took everything. They were [07:51] effectively able to brain drain it like [07:53] Silicon Valley and boom, they have their [07:55] own version. It makes me start thinking [07:57] that maybe something like SQLite is what [07:59] we're going to start seeing. So, SQLite, [08:01] for those that don't know, part of their [08:02] quality assurance is this test harness [08:04] number three. Now, test harness number [08:06] three or TH3 [08:08] is the test suite is a private set of [08:10] test cases used by the SQLite to 100% [08:13] MC/DC [08:15] in an as-delivered configuration. The [08:17] TH3 sources are served on the same [08:19] server as the other SQLite repositories, [08:22] but differ from the others in being [08:24] proprietary. The TH3 code is only [08:27] accessible to SQLite developers. So, [08:31] think about that. SQLite keeps a large [08:34] part of their test private. Now, for a [08:35] long time, did this make a lot of sense? [08:37] Honestly, I don't really understand why [08:39] they did that. That's not the only thing [08:41] they keep private. They also have some [08:42] other things that they keep private, [08:44] such as this libfuzzer. It's to kept [08:45] private, so we do not want hackers [08:47] gaining access to that technology. [08:49] Sounds like I don't know what's going on [08:50] there, but that sounds like Batman level [08:52] going on. And the name of it looks [08:54] just like a typo. Like I mean, that [08:55] looks fantastic. I want whatever's going [08:57] on in there. But anyway, this actually [08:58] has kind of an interesting impact, [09:01] right? Do you think we're going to see [09:02] the day where, say, company-backed open [09:04] source no longer like makes public any [09:07] sort of testing? Because one can imagine [09:09] that if you have a sufficiently large [09:11] project, you can't just simply just say, [09:13] "Hey, AI, go make a copy." It'd be very [09:15] very difficult. Will we just start [09:16] seeing kind of a large change in open [09:18] source? Because what's the purpose of [09:20] producing all this stuff if somebody [09:22] else can just say, "Hey, that's mine [09:23] now." Or even more importantly, your [09:25] competitor can just say, "Oh, hey, [09:27] that's mine now." [09:29] Interesting, at least. Interesting to [09:31] see what's going to happen over the [09:32] course of the next couple years. I think [09:34] we may end up seeing perhaps the end of [09:37] making money off of open source by [09:39] having a company also do something [09:41] special with the open source because [09:42] people so easily can now copy that. It's [09:45] a very interesting future because this [09:46] fork that they have created, I wonder [09:48] how maintainable is it? Is it Are they [09:50] permanently just required to live in the [09:53] slop universe until someone can finally [09:55] wrangle their head around it? because [09:57] the attempt to rewrite Next.js to be [09:59] able to work on other platforms has been [10:01] tried multiple times and it has been a [10:03] failure quite a few times. Cloudflare [10:05] even mentions that they have attempted [10:07] multiple times. And so, what do you do [10:09] with this just steaming hot pile of code [10:12] that you've now been delivered? I don't [10:14] know. There's a lot of just like weird [10:16] world implications in this. We are [10:18] seeing [10:19] truly for the first time a major [10:21] technology that's going to be used by a [10:23] lot of people in which I don't think any [10:26] individual can reason about what it's [10:28] actually doing currently. [10:31] The name. [10:34] Well, [10:40] I don't even have a joke. I don't even I [10:42] I don't even have a joke to make, okay? [10:43] The name's the prime engine. That's it. [10:45] It's all I got for you. A gen. Hey, is [10:48] that HTTP? Get that out of here. That's [10:50] not how we order coffee. We order coffee [10:52] via SSH. terminal.shop. Yeah, you want a [10:55] real experience? You want real coffee? [10:57] You want awesome subscription so you [10:59] never have to remember again? Oh, you [11:01] want exclusive blends with exclusive [11:03] coffee and exclusive content? [music] [11:06] Then check out Kron. You don't know what [11:08] SSH is? [11:10] Well, maybe the coffee's not for you. [11:12] Terminal [singing] coffee [music] [11:15] in hand, [11:17] living the [11:18] dream.  [[3. Resources/Readwise/Full Document Contents/Articles/Open Source Might Change Forever|Full Document Contents]] ## Highlights - Last week, Anthropic released this marketing video in which they're like, "Hey, we built a C compiler from scratch." You know, like we just walked away and two weeks later was like poof, C compiler and everything worked. I could build Linux. It was easy peasy. Engineering's over. Now, of course, this led a bunch of people to kind of actually say, well, hey, it wasn't really from scratch. He used a bunch of test suites and everything. But Cloudflare, when they saw that, they said, well, wait, wait, wait a second. Hold my beer. I got an idea. And they decided they're going to rebuild Nex.js, JS, which is if you don't know anything ([View Highlight](https://read.readwise.io/read/01kjnec5e0838a0kxs2vngxrzh)) - Tags: 👻 ai highlighted - there, not all the things that made uh Next.js big are there. So, as time goes on, size will go up and then they'll become much closer, like within 5% of each other. I think this is probably the most likely answer is that Vinex is just not doing things and that saves them a lot because the only other answer is that Next.js has so thoroughly screwed up. They're sending effectively 2x more than they need to. And honestly, I refuse to believe that. Like, just a no ([View Highlight](https://read.readwise.io/read/01kjnec5fgbzep0mc81zyt5zn9)) - Tags: 👻 ai highlighted