145: Marketing as Engineering  🔗 **Source:** [145: Marketing as Engineering](https://www.youtube.com/watch?v=zpjQLAoLR2w) > [!note]+ Document Notes > > [!info] Generated analysis — not the author's words. Sections below are machine-written > > from the captured source. Only text inside `> [!quote]` blocks is verbatim from the author. > > Do not quote or attribute any other line to the named author. > # 🏷️ Title: 145: Marketing As Engineering > ## 🧭 Table of Contents > - [[#🔗 References]] > - [[#📝 Summary]] > - [[#📖 Main Content]] > - [[#📚 References]] > - [[#🗣️ Paraphrase]] > - [[#📰 Newsletter]] > - [[#✅ Action Items]] > - [[#🧪 Hypotheses]] > - [[#📊 Methodology]] > - [[#❓ Questions]] > - [[#🔮 Future Research]] > - [[#💡 Implications]] > - [[#🧠 Learning Styles Assessment]] > - [[#📊 Accessibility Assessment (WCAG 2.2 AA)]] > - [[#🌍 DEIB Assessment]] > - [[#✅ Logical Validity]] > - [[#🛡️ DARVO Analysis]] > - [[#🎯 Conclusion]] > - [[#🔄 Refactor Appendix — 145: Marketing As Engineering]] > ### 🔗 References > - **Aaron Francis** — co-host, VP of Marketing and Community at Laravel > - **Ian Landsman** — co-host, founder of UserScape/HelpSpot, creator of the "Solo" AI-agent coding tool > - **Sam Sappenfield** — Laravel field-events marketer; built a Laracon photo-sharing app solo with Abby > - **Abby** — Laravel marketer, co-built the Laracon photo app with Sam > - **Emily** — Laravel product marketer; built a "choose your own adventure" prospect-facing slide-deck generator > - **Haley** — Laravel marketer generating ad copy/collateral via agents > - **Chip** — Laravel marketing engineer who finishes marketer-built prototypes (DNS, hosting, path forwarding) > - **Jesse Hanley** — founder of Bento (email marketing/CRM platform), upcoming guest referenced repeatedly > - **Jason Cohen** — WP Engine founder, mentioned re: a new interview podcast with Rob Walling > - **Rob Walling** — mentioned as co-host of Jason Cohen's new podcast > - **Steven Reese Carter** — founder of Valorem Security (episode sponsor), Laravel security auditor > - **Taylor Otwell** — creator of Laravel, mentioned re: Laravel Live Denmark and board composition > - **DHH (David Heinemeier Hansson)** — mentioned as a fireside-chat speaker at Laravel Live Denmark > - **Tom** — president of the Laravel board, mentioned in closing banter > - **Tools/products cited**: Linear, Notion, Grok/"Grockbot", Claude, OpenAI/Codex, Solo (Ian's app), HubSpot, Laravel Cloud, Bento, Dropin Blog, Zvix > ## 📝 Summary > - **What happened**: Episode 145 of the *Mostly Technical* podcast (Ian Landsman and Aaron Francis) opens with vacation banter, runs two sponsor breaks, then covers three main threads: (1) Aaron's account of Laravel's marketing team becoming "engineering-pilled" — non-engineers shipping real production apps and AI-agent-built collateral with minimal engineering support; (2) Ian's viral "DSA" (data structures and algorithms) prompt for orchestrator/sub-agent codebase audits; (3) a Solo app development update (workspaces, command palette, session capture) and travel logistics for Laravel Live Denmark. > - **Why it matters**: The episode is a real-time case study of a marketing org restructuring itself around AI-agent-assisted, self-serve tooling — marketers prototyping to ~95% completion and handing off only the last mile to engineering, rather than filing tickets and waiting. > - **Stakeholders**: Laravel's marketing team (Sam, Abby, Emily, Haley), marketing engineering (Chip and one other), DevRel, the broader Laravel/PHP developer community, and indirectly any DevRel/marketing org considering similar AI-tooling investments. > - **Risks**: Heavy reliance on anecdote rather than measured outcomes; the "just let them build sloppy internal tools" ethos could produce technical debt, security gaps, or shadow-IT sprawl if unchecked; conflates a well-resourced, high-trust team culture with a broadly replicable process; the DSA-prompt segment leans on one person's self-reported success (no code review shown, "I didn't even read through them all"). > - **Bottom line**: A credible, specific account of applying software-engineering norms and AI agents to marketing execution — most valuable as an inspiration case and set of testable patterns, not as a validated methodology. > ## 📖 Main Content > ### 🗒️ Notes > - **~0:00–13:00 (cold open, non-substantive)**: Personal vacation banter about a beach-resort cabana rental; establishes hosts' rapport and disconnect-from-work habits (one host ran "bots" — AI coding agents — remotely while on vacation, checking in a few times a day; describes it as "a week of work done in a day at the pool"). > - **~13:00–16:30**: Sponsor reads (Bento — email/CRM platform; Dropin Blog — headless blog CMS with Laravel SDK) followed by discussion of a new interview podcast by Jason Cohen and Rob Walling, and a lighthearted jab at Jason Cohen questioning why a sponsor advertises to the Laravel audience. > - **~16:30–29:00 — Linear/Notion segment (direct lead-in to the core topic)**: Aaron explains why Laravel's marketing team uses Linear (company-wide adoption, cross-department task assignment, project templates for recurring workflows like webinars) despite Linear's engineering-cycle features not fitting marketing's reactive cadence. Both hosts describe preferring their own MCP-connected coding agents over vendor-native bots (Notion AI, Linear AI) for control and trust reasons, while acknowledging vendor-hosted agents make sense for non-technical staff at scale (e.g., a 50-person help desk) and for scheduled/routine automation. Aaron relays what he frames as a key insight from upcoming guest Jesse Hanley: "take a process that already exists that you're already doing and turn that into bot work. Don't invent it." > - **~29:30–33:00**: Tangent on a "life OS" personal-automation concept (deferred to a future episode with Jesse Hanley), sponsor reads for Valorem Security (Laravel security audits) and Zvix (event-driven webhooks). > - **~33:00–45:00 — "Marketing As Engineering" segment (core topic)**: Aaron walks through concrete internal cases: > - Sam Sappenfield and Abby built a full Laracon photo-management/sharing app (upload, rotate, resize, background image-processing jobs) without involving product engineering; marketing engineering's only contribution was DNS/path routing to plug it into the main Laravel site. > - Framed as "the dream": a marketer gets an idea 95% of the way to done, then marketing engineering "pushes it over the line." > - Emily (product marketing) built a "choose your own adventure" slide-deck generator for prospect conversations, mapping stated customer problems to supporting material. > - Haley uses agents to generate ad copy/collateral variations from a set of design-approved templates. > - DevRel team members are building autonomy tools too — e.g., automated thumbnail/OG-image generation so they don't have to pull a designer off higher-priority work for small assets. > - Aaron's explicit management stance: internal tools can be arbitrarily "sloppy" under the hood — only the output quality and speed gain matter ("Does it make YouTube thumbnails? Does it make you faster? Then let's go."). > - Both hosts connect this to Sam's viral tweet thread (documenting her build-in-public learning process), reading it as an example of authentic, journey-based content outperforming "aphorism" style thought-leader content (with a pointed dig at James Clear/Naval-style tweets), and as strengthening the "artisan builder" culture of the Laravel community welcoming a new non-engineer builder. > - **~45:00–55:00 — Ian's viral "DSA" prompt**: Ian describes a long-standing personal prompt ("DSA" = data structures and algorithms) that asks an agent to find material simplifications after a coding session (collapsing scattered booleans into enums, promoting shared assumptions into first-class objects, etc. — changes that "eliminate the possibility" of certain bugs rather than fixing individual instances). He recently extended it into a two-stage orchestrator pattern: an orchestrator first catalogs all subsystems in a codebase, then dispatches a fresh sub-agent with the DSA prompt into each subsystem, then the orchestrator deduplicates/ranks findings into a scratchpad, and finally a second pass turns the scratchpad into a discrete, linked to-do graph that lower-effort agents execute overnight one by one. He reports the Hello. >> Back from the cabana. You look You look recharged. How do you feel? >> I slept bad last night, but besides that, I am recharged overall. I uh got the sun was all over me. Got the cabana. We should discuss. >> You love a cabana. >> A cabana. So, I got a lot of cabana thoughts. >> It's going to be a long one. Um here's the thing. We're at this resort. It's like the Sand Pearl Resort. It's like one of the nicer, maybe the nicest resort in Clear Water, Florida. Okay. And it's like, I don't know, it's like $500 a night for a room or whatever. $600, something like that. They have five cabanas. We had the only cabana. We had a cabana every day. We were there for 4 days. And we were the only people with a cabana the entire time. And I don't understand why. Like >> so without getting without getting into family business. How much is a cabana in relation to a room? >> So with the cabana was $500 >> a day. >> Yes. And I can understand not doing it your whole We went a little overboard. Maybe you don't need it every day, but to have it no days. I mean it's hot. It was hot. There was heat advisories. It's like you're spending $2,000 minimum if you're there most likely plus all this other stuff. I mean 500 bucks. You're in a cabana. >> And what what is a cabana? What does a cabana afford you besides shade and a place to like camp? What do they get? >> So you get you get So you got a Yes. an enclosed space. Is that in this case actually part of a building which I hadn't seen it quite this way before which was quite nice. You got a fan. You got a television. >> Love a fan. >> You got a refrigerator. You got a guy who comes the cabana boy who comes and sees if you need anything. You got a >> Then you pay for it, right? It's not like5 is it $500 to get the spot or is it 500 plus some food? So, it's the spot, but you do get waters to start and you get a bowl of fruit bowl things. You know, there's a little bit of stuff. Any other food you do pay for. Uh but but that's true. Even if you were like just in the chairs by the pool and you wanted to eat, you would pay for the food. So, that's fine. Um and what else you got? You got you got four chairs dedicated to you outside. So, if you want a sun or whatever. Yeah. Well, they're like in front of the cabana, but it's not Yeah. At the pool, but in front of the cabana. So, I mean, all the chairs were packed every day. So, like, you know, it's a whole thing. You can't leave your chair more than an hour. So, whereas the the four in front of the cabana, you do whatever you want. You just you sit in them. You don't sit in them. Nobody's going to come take your chair. It's all for you. I'm just saying, everybody there looked like they had a few bucks. Like, nobody not a single person in this big hotel resort was like, "You know what? Look at these this family. I see this family in this cabana [laughter] every day. They look like they're having a freaking great time. Let me get one of those cabanas. They look normal. What am I missing out on? They can do it. >> Meanwhile, I'm telling my kids how I'm getting a cabana life tattoo across my chest. That's what I'm doing. Cabana life. >> Plus, plus you get to be royalty, which is half the fun, right? You get to you get to look at the chairs and say, "I don't want to use those." And that's half that's half the fun, honestly. >> Yes. There's a cabana boys coming to wait on me. Go out to the beach. You come back. You got a nice area all to yourself. If you can relax, you can lay on the couch. You could take a nap. There's a We didn't actually do this, but there is a uh cover so you can have pure privacy. If you want to be in total privacy and take a nap, you can do that. Um, yeah, it was great. There's two chairs also. I mean, there's like seating for like eight, 10 people in there. It was fantastic. So, I don't know what these people are thinking. This makes your vacation. If you have the option to get a cabana, just get the cabana. At least for one day. At least one day. Get the cabana. >> Always get the cabana. I think I I actually You know, we go back on whether or not you should do Laravel new on outro again. On this advice, on this advice, I'm I'm on board. Always get the cabana. >> Always get the cabana. >> I mean, there's a lot of people there who are obviously there with other families even. So, it's like a >> That's an even better use case. >> Everybody puts in 200 bucks and you have a cabana and the kids can go crazy and they have a base of operations. You can leave your stuff. Oh, you got your iPad there. Just leave it in the cabana. You don't have to like hide it and try to figure out what you're going to do with it when you go out to the beach, whatever. Just leave it in the ground. >> I love a base of operation. >> Oh, it's so key. >> I love having a spot. I love a spot. >> This is a spot. >> We used to always go to Martha's Vineyard. We've been off that the last few years, but my big thing with Martha Vineyard, you have the truck and the truck is the base of operations. I just put all the in the back of the truck. You drive the truck onto the beach and you and your are right there as the base of operations and it's fantastic. And so you can't do that at the resort like this, but the cabana is the truck. The cabana is the base of operations. >> Bring as much stuff down as you want. Bring all the toys you want. Do whatever you got to do. I'm a little past the toy phase at this point, but like you just have all your in there. >> Especially when you're in the toy phase. >> This makes a thousand% sense. >> Toy phase is the worst. You're freaking changing diapers on little babies. You want to be doing that in a cabana. >> You got the little You got the little potty thing so they could do the potty when they, you know, you probably have a couple of those little potties around. >> So many so many We walked out of church yesterday. They've been in church for an hour and a half, two hours, whatever. We walk out, we get in the car, one of my children says, "I need to use the potty now." And I'm like, "Why [clears throat] now?" We were just in this. >> We've literally been in church for 2 hours when there are bathrooms everywhere. So, I had to get the car potty out. Hate the car. >> Car potty. Hate the car. >> Jamie loved the car potty. Like, hate it. >> The best one. Do you have the one? The best one is the one that's a Ziploc bag. Do you have the Ziploc bag one? >> We have the one. We have the one where the bags like have a tying mechanism built in, but not not a ziplo, >> but there's some kind of bag situation. >> Yeah. Yeah. Yeah. Yeah. >> Okay. You got to have a bag situation cuz then you just toss toss the stuff. You don't want to be cleaning out. >> Just throw it out the window down the highway. >> Yeah. >> Did you uh did you fully disconnect? I know that you you did a service to the community by doing the pod, but how how much are we disconnected when we're when we're on vacation? >> So, I mean, for me, this was fairly disconnected. Um team took care of the day-to-day stuff, so I didn't have like any like day-to-day things, which is of course the best. I love that. >> And also, just to clarify the record, Jamie works in the business pretty hands-on, too. So, you're both you're both on vacation. And did you both disconnect? >> We both disconnected. Her stuff for the most part, she can push a week and it's not too bad cuz it's like she mostly does like accounts payable, accounts receivable, uh chasing people for money and like and then the books and all that stuff and then like other operational things, insurance and whatever, all that stuff. So, >> most of the time it's okay. Yeah, I think she did check in a few times just to like if there was like somebody had anything urgent, but yeah, she mostly is disconnected. Um, and I mostly disconnected other than the uh the bots. I had the bots going on a HubSpot project, which I talked about last week. And oh man, it's so cool cuz like in the old days I would go on vacation, but I would do crazy Like I remember one year we were in Disney and I was interviewing people to hire for a job and it was like this is crazy. like I'm up in the room doing interviews on vacation like this is yeah terrible. Um >> so and then like coding and whatever but this I got the coding done. It's like oh in the morning I kick off the bots like before dinner time we end up back up there. I check in on the bots and uh at night before bed I might kick off a bot and that's it. And it's just like they're they're rolling. They're doing >> a week a week of work done in a day at the pool. >> Yes. Exactly. I didn't even bring the laptop to the pool. like the laptop just stayed in the in the room and I just checked on it at those couple times a day. So, I was still working. There was no total rest for the weary, but it felt quite relaxing. Um, so overall, I would say it was like 95% of a uh of a disconnect. So, that was >> that's pretty good as a business owner. 95% is pretty wild. >> Yeah, I like that. We we also went we normally do like at least a week. Sometimes we do more than a week. Um, but this time we did just it was like I guess it was five nights. Um, which I thought was good for like Florida. There's like less stuff I want to do, just like stuff around. So, it was like mostly we just did the beach. So, we're just at the beach and uh that was enough of just like pure beach time. I'd never been to the Gulf of Mexico before. That was cool. Um, >> I don't know if that's real. Isn't it the Gulf of America? >> Yeah, it is. So, there's a So, there's a thing called >> Make sure we're on the same page. >> There's this thing called surf style. Has this little jingle surf style. Surf style. If anyone's been there. Yeah. Yeah, it's a good jingle. >> Yeah, >> I did take this picture. I was going to post it, but I was like, whatever. I'm not gonna I don't even want to I don't want to put the negative energy out there. So, I >> go, Ian. Hey, [screaming] that's But um it was quite funny. It was a little on the nose. So, the surf style place is what you might imagine. It's like beach gear clothes, whatever. Big place. And then they have a section Gulf of America shirts, right? And other little trinkets. little little stickers or magnets of Trump, whatever, all the stuff. Big red sign, huge red sign. Buy buy one get three free. [laughter] I was like, you know, this is just so on the nose about like America. These were 100% made in China. Um the whole thing was just right on the nose and I was like, this is just all too much for me. But anyway, >> yeah. So, and the first two days were like a bathtub. So, that was cool. It was like cuz Martha's Vineyard is like you are just getting pounded. It's like boom, waves, big old waves attack. The kids can't make it more than like 4 feet out there. And like that's >> Is it like a rocky beach at Martha's Vineyard? >> Um, some of them are. The one we don't really go to. It is definitely rockier than this was. This was like no rocks at all. Um, but it's not like the huge boulder type rocks, but there will be like some parts of goodsized rocks and yeah, so the rocks are hitting. I mean, the kids always come out bloodied, bruised. Like, that's like [laughter] that's like what going in the ocean is in my vineyard. Whereas here, it's just like you're floating around. There's nothing in the water. The there's no waves. You're just literally in this the world's biggest bathtub. Um the back couple days, there got to be some more waves. They were still not uh ocean level of wave, but um it's a nice little change up, but yeah. >> Well, welcome back. Thank you for uh thank you for giving the people what they want, which is an episode on location. >> Yeah, that was so much fun. That was cool. We got to experience the cabana. You could do a podcast in your cabana. This is the point. >> So, it's a writeoff. >> You you performed business inside the cabana that you could not have performed at the pool. And so, >> I didn't actually I actually paid for it myself, but now that one day I should write off that one day. Yes, that's true. And if not one, if one, why not four? That's what I'm saying. >> That's what I'm saying. That you're an accountant. >> This is This is why I'm a former accountant, by [laughter] the way. >> No, >> cuz I just Yeah, that sounds good to me. >> How does that work with the CPI? You You keep it forever. You have to maintain. >> I kept it for a while. I kept it for a while. It is now I think it's now fully lapsed. You can go inactive without keeping up your continuing education, which is what I did. >> Oh, that's right. The continuing education. And then at some point I was like, I don't even want to pay the the money every year and file the form every year. Just it just doesn't matter. It's getting me nothing. >> Yeah. You're not you're not going >> because because legally I would have to say if I were to ever sign my name anyway as CPA, I'd have to put inactive in parenthesis. And I'm like >> I'm not I'm never going to sign it first of all. Second of all, I'm not going to put inactive next to [laughter] it. >> It's just a better It's a better story at this point to say I was a CPA than to actually have the license, you know? >> Yeah. Yeah. And the continuing ed is like kind of real if I recall. >> If you if you want to actually keep it, it's kind of real. Yeah. >> Yeah. You can't just do like, well, I'll spend one day every couple years and bang it out. It's like it's like a whole thing. Yeah. >> Yeah. >> So, um Okay. So, no accountancy for you. We're not going to we're not going to count that. >> No, I think I think we're out of uh we're out of that window. Never going to go back, I'm afraid. >> Hopefully not. So, how's we have a we have a couple things to talk about, but you know, week two, just give us a little update. What's going on? How are things? >> Week two is brought to you by Bento. >> Okay, >> Bento is the email marketing and CRM platform built for the AI era. Send your product marketing and transactional email with Bento. And check out their new transactional email pricing with rates as low as 9 cents per thousand emails. You can find them at bento.com. I think we have a sweet hang with Bento founder Jesse coming soon very soon. >> Uh perhaps even tonight we'll see. And then also brought to you by dropin blog. Dropin blog is a headless blog CMS with a firstass Laravel SDK. So you get your domain, your design, and your stack at dropinblog.com. I know you've been traveling, but did you see that the drop-in blog founder was on with Jason Cohen and Rob Walling? >> I did. I didn't get to listen to it yet, but I it's on my list here of uh pods to listen to. So I is this is a new podcast they just started. >> This is a new Yeah, this is a new ad hoc thing. J Jason's done uh he's done two episodes now. This is with with uh actually Jesse. Jesse from Dropin Blog was the second episode. Um I'm about halfway through it. Um and it's it's very very good. It's a little adversarial for me. >> Oh wow. Oh baby. Let's go. All right. >> I I I feel like I feel like Jason is uh is really like pushing hard and sometimes I'm very uncomfortable. [laughter] I'm like, "Oh, let the guy finish. Oh, this is making me nervous." >> No, like a podcasting etiquette angle or just in terms of business knowledge. >> A little bit of a podcasting etiquette angle. I think I think Jason is like trying to keep the show moving and like dig in on in into people's assumptions because people will people will say things and Jason's like, "What what do you mean by that?" No, that doesn't make any sense. Um, and that part I appreciate. It's just the like I feel like there's so much tension in the room. Um, [laughter] and some of that some of that is stage craft of the podcast. So, it makes me a little bit nervous, but >> Oh, man. >> Uh, they're good episodes. I would recommend them. Well, I'm happy to see because like Jason I feel like is one of these guys that got so much knowledge and uh OG status and everything, but I feel like >> uh he's got like I feel like a podcast is good for him to get into. Like he's got to find his fit in the new world and I think like just tweets that are very smart and logical like that's not what Twitter wants. They don't want like here's some good advice, right? So, like I feel like that that it clicks with me and I always like his stuff, but I don't know. Doesn't seem like it totally clicks given his stature and knowledge level. I feel like you got to give it the podcast a better angle. >> He he's one of very few people on Twitter who has less uh like hype and credibility than he deserves. It's like he deserves he deserves to be like the foremost thought leader like on our side of Twitter >> and all these other people who are the foremost thought leaders maybe should take a step back. >> Um because Jason Jason has the lived experience that has earned him this kind of stuff multiple times. >> I'm happy to see him in this new medium. Uh so we'll we'll add a a link down below for uh the drop in blog episode. So, but it's very good. >> We're in the circles. Our people are out there. >> I know. >> Yeah. >> They Jason, my only bone to pick is Jason told him like, "Why are you advertising to Laravel people? You don't need to be advertising to Laravel people." AND I WAS LIKE, "OH, JASON, >> did they actually say that? >> Food on my family." Yes. He actually talked about it. >> Oh, baby. Okay. We got >> Yeah. Jesse was like, "Yeah, we're advertising on Mostly Technical. We went to Laracon." And Jason's like, "Why are you doing that? Don't do that." I'm like, "No. [laughter] All right, forget everything I said. Guy's an idiot, obviously. Like, >> that's what I'm saying. Unbelievable. >> All right, I got to listen to this now. >> Second week. Second week has been good. >> Okay, we're still throwing haymakers. What's going on? >> Uh, yeah, we're still we're we're throwing a few hay makers. Um, I think things have settled down quite a bit. Um, I feel like the team, >> there was never a question of if the team is doing the job, but the question is like, do we have the processes to support the team doing their jobs. >> Um, and I feel like that has settled in just a little bit. We still have we still have a ways to go, particularly with like figuring out how the hell to use linear appropriately. Um, >> which there's Yeah, there's question there's no right answer. Well, okay. Let's dig into that. Let's >> dig the right answer. >> Let's dig into that. We do have a segment later where we're going to talk about how you are basically just turning marketing into engineering. So, that that's just put that aside for [laughter] a second. >> Um, why is marketing using linear? Like, is Linear even the right tool for the marketing group? How are they using it? Now I have because you talk about linear and I'm like I don't I kind of get what you're using it for I think in theory but at the same time it seems like there's it brings so much other stuff that marketing is just never going to use that feels like maybe not the best fit. So >> yeah, give us the layout. >> So I think I think linear is good first of all as far as like base case it's way better than my base case with notion for example. >> Linear linear is a very good tool. >> I think uh in some regards linear has become all things to all people. And so they've got this they've got all these these concepts in there that just simply do not apply to us whatsoever. So they've got like life cycles and triage and SLA and initiatives and I'm like what we don't need any of this crap. Um >> but why do we use it? >> Um you can imagine a scenario where like we're running a webinar which happens very often. That process is pretty pretty baked and so you have like we'll say six to eight things you need to do for every webinar. That's already a good case for lineers like templates, project templates. So, it's time to do a new webinar, spin it up, it populates, we need to get, you know, the slides done, the people, the copy written, the thing on the website, yada yada yada. >> Um, for each of those things, it might be a crosscutting concern, right? You might need to talk to design to get uh some assets made. You might need to talk to marketing engineering to get the website updated. you might need to talk to engineering to get like the content ready. >> And so part of it is that's where the whole company works. And so being able to go in there and tag the person who's giving the webinar and say like, "Hey, here's your task in linear fits within their like life cycle of day-to-day work, >> right? >> So it's a good it's a good coordination mechanism. Okay? So that like uh Anna who does our webinars, she can go in and create the template and then assign Emily the email copy and Chip the website updates and all of that stuff. So that part's really nice. >> Yeah. >> Then you've got stuff like we need to publish a blog. We need to put out a YouTube video. We need to send we need to send an email for uh Laravel Live Denmark. And all of those things are discrete deliverables with due dates that likely include sign off or collaboration from other people. Okay. >> So that fits a that fits a companywide task manager pretty well. The problem is I think like for example engineering works in cycles which is very like shape up base camp style. >> Marketing doesn't work in cycles. like we're just we have to be too reactive and we're not like we don't have you know a year's worth of road map of stuff we got to do. We're kind of just reacting to product and the cultural millu stuff like that. And so some of the concepts inside of linear just don't apply. But on the whole it's nice to be able to have somebody from outside of marketing drop something on our linear board and then we can like assign it internally, assess it, set a due date, etc. So if it's getting used cross department then I think it makes a lot more sense because if you guys were often like to-d doist or sauna or whatever then like when you had stuff for other departments like you'd have to figure out how to get it over there all that kind of stuff. Yeah. Okay. >> And that's a big part of like that's a that's a big part of like my thought process is >> why are we using linear? Why are we using notion? because the company uses linear and notion and like we're just going to make it work for us. These tools are not bad. I think they were just poorly maintained and so they're a little bit of like it's a little bit confusing just because of all this old like historical artifacts that are in there. >> Um and even already with notion I'm seeing like I actually don't hate it that much anymore because I've stripped back the way that we use it. And so none of these tools are bad. It's just like these are the tools we're using. And as much as I want to like burst through the wall like the Kool-Aid Man, I'm not going to change a companywide tool. I don't even like I don't super want to, but I super don't think that's a good use of time. And so we're just we use the tools we have and we're going to use them better. And I think that's a good like that's a good place to land. >> Yeah, I like that. And then also like I mean you could imagine a world where like Linear's uh MCP is actually really good. So like you could be doing stuff with that for people who maybe don't need the full power and it's like overwhelming. There might be other paths to letting them interact with it. >> That's how I'm using these things. I mean I I'll go I'll go into solo with codeex and be like hey what the hell's going on in linear like why does this feel so disorganized and it just reads through the MCP and it's like well it feels disorganized because you've got this this and this and I recommend this. And I'm like thank god you're here you know. >> Exactly. So we do have the tools now that make it uh easier to deal with some of these complicated systems if you don't want to deal with the or don't need the complication necessarily all the time. So >> one one interesting like revealed behavior in me and I'll be curious how it is for you. I don't use the notion agent nor do I use the linear agent. I use my own agent hooked up through MCP and I think I don't know what it I don't know like where that revealed preference comes from. Part of me is like, I don't want to pay an unknown amount of dollars for an unknown amount of tokens inside of notion. Like, I'm good. I I've paid already. >> The other part is like I don't know if I believe that their agents are very good. Like, [laughter] >> surely surely they're not running 5.6 soul extra high when I'm [clears throat] paying them a nickel, right? And so, I I don't want to use their agents. I just use my own agents for these tools. Is that the case for you? >> Yeah. Yep. That's uh I don't ever use their internal agent. Um but see now this is interesting because I think we might end up talking to Jesse Hanley a little bit about this cuz something along these lines came up. I do think though that technical people take that to mean that the having an agent in your app is stupid and I don't think that's true at all because the >> I don't know if I go that far. >> Yeah. because it's like for us it's stupid because we are like super dialed in and I have 10 ways right here to interact with my own agent in my own setting however I want to do it and that's just not going to be true of we'll take the example I know you have a help desk with 50 people all 50 people don't have their own claude all 50 people even if they have their own claw or not hooked it up to the MCB like they're just not working that way now >> I don't think they're going to be working that way ever probably or at least not for a long time. So, so those staff still need the agent built in. And for them also, I think the part you're talking about is also less of an issue because they're not asking it to do things that require soul extra high to think for >> analyze my entire workspace and come up with a better system. Yeah. >> Yeah. And even that nowadays can be done by a lesser model to some degree. Like we don't want to deal with that. We just want just want the best model. >> That Yeah. But I think that and the vast majority of time is nothing like that, right? It's like, "Hey, like like you just did it with the show before we started. You used Outros Gopher and you said, "Hey, add this thing. Hey, add that thing." And it added this thing and added that thing, right? Like it's just like straightforward stuff. It's just going to do it. Doesn't matter that it's running Sonnet 4 or whatever cuz I haven't even moved it up yet probably. like it just made a thing in the list and it's not like so you know it's like stuff like that where it's like the the average worker be is not uh going deep on yes reorganizing their entire structure of their system or whatever like it's just not the work they're doing. So >> yeah, I think one place [clears throat] one place that um these internal like first party hosted agents could win or could be more helpful is the routines, the schedules, that kind of stuff. So like in notion you can set up like every morning at 8 I want you to like do this thing. Who knows what impossible to say like what I need a daily notion thing for you. >> But if I did uh having that hosted on their side and their infrastructure with and and I just know that like after I set it up it runs. I could see that being a pretty big use case. And then also like the team aspect of an agent. You might set up an agent that is inside of notion or linear because in notion you can make like kind of like Grockbot you can make like named agents or whatever and I can see that being useful from like a team sharing perspective. So I don't think it's a I don't think it's dumb to have the agents. I agree with you but I'm just not using them. I just find the MCP is gives me better like control. >> Yeah. Yeah. And it also depends. I think there's like all these different levels people just like dump it all in as one thing when it's like not one thing at all. like like even these these bots like bot grobbot or a notionbot like there's certain things it can do on its own but like there's things where people like they're going to use it for everything and we're probably going to also talk about this with Jesse but like I don't even know the answer to the thing that is the question like something question comes up if I don't know the answer I can't how am I going to tell the bot to do the thing I would do because I don't even know what I would do yet and I have to >> therein lies the whole struggle of the modern age >> exactly [laughter] >> you got to know what you want. >> Exactly. So, it's like that I'm going to need to think about it myself. So, I can't automate that part yet. Or there's like I think the bots are doing more advanced things and business processes are interesting and that's something that and that's like what we're building in HubSpot and you're not going to do that from your local machine. I mean again some people are some people are like I set up claw and it answers my support and that's fine on like for a very small type of setup but like the HubSpot one is like 20 point check. It's got all these checks and rechecks and some of them are AI and some of them are uh deterministic and like there's all these steps in the process for like a more sophisticated and like professional let's say flow of a formal help desk that's got a bunch of people and has all these rules and things they want done a certain way or whatever, right? It's not just like, hey, I'm me and I'm just like answering the outro support with AI. Yes, great. Read the docs, answer the support. Awesome. Like it's like more complicated than that. So that's going to be again something that you're not going to be able to easily set up on your own through an MCP like >> agreed. There's just more to it, right? And so again, not impossible to do probably if you really wanted to architect the thing out, but I don't think most businesses want to do that. It's getting too complicated. So then yes, you'll have an internal agent that has all this architecture behind it that does a business process that makes sense to have as an agent. Like at least when I last used Notion, it wasn't that. I know it's moved beyond. It's more sophisticated now, but it was very much the like, hey, ask me a question about your data type of AI thing. So, you know, but that's what that's interesting about these generic tools is the bots I think are very tricky cuz like what do you want the notion bot to do? Like there isn't like a form a business process everybody's doing, right? I have no idea. >> My help desk app I know there's basically one business process this whole app does and let's just have the bot do that business process and there's more to it than that, whatever. But like that's the guiding star, you know? Whereas like notion is like I don't know it could do anything. Like what should it do? >> Who freaking knows? And we can't we can't spend the whole episode talking about what we're going to talk about with Jesse. But I feel like this was [laughter] one of his one of his better points is like >> take a take a process that already exists that you're already doing and turn that into bot work. Don't invent it. Aresh. Um you had a tweet about creating a life OS. Should we save that for the sweet hang because that dovetailed so nicely. I think I think we should I think we should >> talk about life west on the sweet hang. All right. >> Yeah. Um because we're going to do it's tonight, right? That we're >> recording tonight. >> So we're recording with Jesse Hanley tonight. He's a big Grockbot guy. Um I'm skeptical. Where do you land? Are you kind of in the middle? >> Uh I think it's again I don't know what I'm going to really have it do. There's some things I can have do, but again the life I guess we could talk about there. It's more like first of all I'm not going to pay Elon any money I not to pay him. So I'm just going to wait for I mean Claude and OpenAI are going to have this within a month for sure within a week probably. I mean they already have the basic structure. It's like just like the kind of UI. I don't know. There's something weird about how they both do server stuff that I don't totally understand. Um so I think there's some stuff to work out there which is like mostly UI and maybe some infrastructure but uh whatever. They're going to have this stuff and I guess I'll throw some stuff at it but I don't know. I don't feel like I have that much to automate purely through a bot. Like I just think there's more. I need UIs. I need other things besides just I told the bot to do something. >> And without you know talking about it because we're saving this for the [laughter] next show. I do think >> what I read in your tweet is you want a like deterministic control plane, not necessarily like a free form bot like that. When I hear life OS, I'm like, I want a command center where I can suck everything in and look at it and then like maybe dispatch actions or like take take deterministic uh routines or flows or whatever. So, >> yeah, manual flows or whatever. Like, I want the AI to organize my mail cuz I'm so I can't believe Gmail hasn't given us like they just released this in AI inbox thing. It's freaking terrible. Like, it's totally useless. So, whatever. So, I wouldn't mind organizing stuff. There's things that could auto close blah blah blah, whatever. There's stuff like that and the bot will presumably do the organizing and some uh auto cleaning stuff, but then I need to see the rest of it like and I don't want to have to be like bot make a list and it goes off. I just want to go I want all that stuff to have occurred and like there's a nice UI that's the same every time. Not like the the AI generates it every time and it's a little different every time. Like I just want to have it the same every time in the same spot in the same way. That's what I want. So, you know, we'll see. Jesse, you got your work cut out for you battle. >> Why don't you tell us about Valoran security? >> Oh, you know, I'm so off my game. You're It's a good thing you're here. It is a good thing for me. [laughter] >> So, security Laravel security audits and penetration tests. AI doesn't help you ship code faster. It helps attackers build tools and find vulnerabilities faster, too. That's why you want a friendly hacker like Steven Reese Carter on your side catching the holes before they do. Check them out at valorenssecurity.com. OG Laravel. I know he's been around forever. Like literally forever. I don't know how long, but long time. Super Laravel, knowledgeable. It's done security audits for practically everybody. So, uh, check him out. >> Spoken a couple Laracons, too. >> Yep. Spoken at Laracon. Lar online. Um, uh, Zix, never get paged again from a broken web hook. Become event driven with Vix. Check them out at link.zvix.com/mostlyte. Thanks to everybody for sponsoring. Love our advertisers. They make the show possible. >> They in fact do. All right, we [clears throat] got a couple directions we can go. >> Yeah, >> I had a I had a hit tweet. We can talk about that. >> Oh, yes. >> Marketing is becoming engineering. >> We could we could talk about that. >> Uh maybe we'll end with I'm going to Denmark. So, let's keep that one off. So, what do you want? Do you want marketing as becoming engineering or do you want the hit tweet? >> No, let's do let's do marketing as uh engineering or becoming engineering. Really? >> So, we kick this off. >> Oh man, it's great. It feels good, man. It feels good. >> The influence is already being felt. Look at this. Um, [snorts] so over the weekend, uh, our very own Sam Sappenfield started to get a lot of good, uh, a lot of good Twitter action moving. >> And the reason is is because she's been she's been engineer pilled, she's been AI pilled. Um, so who knows what she's building, you know, on the weekends, which is just the freaking best. She's like, I want to build stuff on the weekends. We got her. It's over. The hooks are in her. >> U, but if you back up just a little bit, Sam and Abby um built a full Laracon like photo management sharing application that will be released this week. >> And so all of the Laracon US attendees will be able to go to this page and like view all the photos. So cuz we had, you know, a couple different photographers running around. Um, >> and this this uh app that they built gives them the ability to like upload, rotate, resize, all that stuff. So, it's a it's a lightweight app, but it's a real app, right? You got to be able to upload and do background jobs for image processing, all of that. One big thing that I am trying to one like uh institutional thing I'm trying to institute is >> becoming more as a marketing organization, not as individuals, but they kind of go together. becoming more autonomous in uh the things that we're able to do. So Sam and Abby were able to build this whole thing out without involving, you know, uh, product engineering and then without involving our marketing engineers either cuz we have two marketing [clears throat] engineers and the only thing they needed marketing engineering for was basically like let's pull this into the Laravel or and then let's get the DNS and the paths the path forwarding all set up correctly. >> Beyond that, Sam and Abby built the whole thing by themselves. >> Wow. Okay, >> which is the dream, right? The dream is the marketer has a good idea. >> They're able to get it 95% of the way there. >> They go to marketing engineering and say, "I'm almost there. Could you push it over the line?" Chip's like, "Sure, no problem. Let me push it over the line and we're done." And so between [clears throat] So like that's that's field events. So that's Sam and Abby. Emily is in product marketing and she DM'd me the other day and was like, "Hey, I want to show you something. that I've been vibing on and I'm like, "Yeah, all day every day. Show me >> spreading." >> Um, and she was working on this this like she's super good at like translating like products into which I guess is what a product marketer does, but translating products into like stories that matter for end users. And so she showed me this like choose your own adventure kind of like basically slide deck generator that we can use to talk to prospects and be like you said this was your biggest problem. Here's our supporting material on that. Um and so it's just like enablement stuff. Then Haley is ripping ad copy or like um uh collateral via agents. We have this like we've got this set of, you know, design blessed ads and then she's pointing her agents at it to make more. >> Yeah. >> And then I've got the devrils working on stuff to to like make themselves more autonomous. And one one way that surfaces is getting thumbnails made and getting OG images made. Like it has to be done. It's not the biggest thing in the world, but you then have to bother a designer and it's like h what are we like they're they're busy redesigning everything else. Do we really want to bother them with a OG thumbnail for a you know forge change log or change log? Not not really. >> Yeah. >> And so um I think they were all I I don't want to take full credit for it. I think they were all pretty like AI curious or at least AI dabblers. But this push to build more internal tooling and build more things that can help us do our jobs better faster. Um I think that's really working. And one one thing I've continued to communicate to them is like build as many internal tools as you want. Like make the internal tools. I don't care if the internal tool is is sloppy. Like I just I just don't care if it produces a good final asset. That's the deliverable, >> right? >> So I don't I don't care if your your uh YouTube thumbnail generator is just a total mess under the hood. Does it make YouTube thumbnails? Does it make you faster? Then let's go. Move on. Let's be done with it. Um and I feel like that is I feel like people are excited about that, but more importantly, um it makes us be able to move faster. So, it's just awesome. I'm having the best time. >> Well, I think the other really kind of key aspect I think is really cool about that, especially with like what Sam and Abby have done in wherever they're especially building on top of like actual Laravel elements is like it's like a Apple. They don't go to marketing and be like, "Here's a new iPhone. Look at it. See it over here." And then like walk away with it and be like, "Market this thing." Right? Like I'm sure they give everybody an iPhone and they're like, "Here you go. Here's a new iPhone. Totally. Let's play around with it. Let's see what the new gizmo does. Here's the new button. go live with the new button for a week and like explore it and like feel the space. >> Fall in love with this. Yeah. >> Yes. So like then you can actually talk about it from a position of knowledge and having actually used it as opposed to just like someone sent you a spec sheet and you are attempting to you know describe the spec sheet or whatever. So >> yeah. So I think the more that the marketers are like hands-on even if it's you know simple tools and basic things and have to be they're building like shipping production paid for applications right it's like that's not your role but the role is like get used to the tool and how it works understand how things connect and like all that it's going to be just super beneficial for for their job for sure and they're going to do a better job so I love that. That's great. >> Yeah. And Sam is like a good uh a good archetype of the future of development I think. Yeah, >> it's like I have a problem. I want to fix it. How do we do that? Um, and so I feel like her insights here are doubly valuable because these are the people that we want to bring into the fold. >> Um, and it's been really encouraging on my side to see because we've we've talked about like the artisan builder gradient, right? We're we're our our history is full of artisans and I don't ever myself being one of them, I don't want to lose those, but I do want to bring in more builders. And if you just go look at like the replies to Sam's >> like tweets over the weekend, >> you see all of the artisans welcoming in a new builder and it's like this is this is exactly what I was hoping for. It's like all the OGs are not saying, "Wow, you just showed up off the street." It's all the OG's saying, "Oh, totally." and then you could also do this or look out for this or hey let me know if you have any questions I'd love to hop on a call and I'm like this is this is easy we've got the best community in the world and we've got this new like wave of developers and turns out everybody's happy so that part was super cool >> just for people who haven't aren't totally aware like she tweeted that she was building this and then she was has been tweeting through the process of like figuring out Laravel and doing the AI stuff and whatever envir environmental variables and like all the steps right that you have to do >> and >> yeah and it's spread even beyond the laral people like you mentioned there's a cloudflare guy chiming in and like whatever there's a bunch of people chiming in and I think it's a great example of like the power of an interesting story which is like what you're trying to tell right is like look how much attention this has gotten like I think Sam said she like tripled her following right and it's like cuz it's an interesting story people are engaged with it people get it like the people who've been there get it and are want to be helpful. >> People who maybe are aren't technical see her doing it and are inspired, right? So like you have this whole thing forming around it and it's like how do you mark how do you get all your marketing to be that way where it inspires people and uh you know is an humans love stories, right? Like here's a story and this is the story that's the hero's journey. >> It's the hero's journey. It's all we're not reinventing what people like, right? people like the same thing forever. Like caveman had the same thing. So it's like how do you leverage that and and reach people through the mechanisms that they enjoy which is actually fairly standard of across people across cultures like it people are grabbed by the same types of stories and so yeah how do you bring more of those to light >> and I I think the generally instructive part for the dear listener is that the stuff that Sam is tweeting is like the platonic ideal of Twitter content. [laughter] >> It's It's not It's not like I did this thing. You should do this thing. I'm a genius. >> It's hey, I'm I'm learning this and I'm going to share it. I did this and it felt cool. I did I learned this thing and I didn't know that and that was frustrating. Like that is what in my opinion the best x.com content is is like following whatever journey you're on following that journey instead of trying to pretend to be a naval and like tweet these pathy apherisms like James Clear and Naval. I'm to I've totally muted them like 5 years ago cuz everything is like what are you even saying? None of this is relevant to anything. Um, so >> it's very authentic, >> very authentic, >> a very authentic journey. She's >> it's a good it's a good case study. So if you're looking if you're looking to grow uh your Twitter following, just go see what Sam Sappenfield has done and just like try to really read like the story and the tone of it. Um, and that I think is that's your north star. So that's been super fun to watch. And of course she's got side projects now, which we love cuz that like >> we love gets her in the game. It's putting her on Laravel Cloud. It's like she's learning all this Laravel stuff. Um, so I'm just thrilled to death about it. >> Then there's no way that doesn't help her throw better events. Like it just has to, right? Like even just in her interactions with people like she's going to know what somebody's talking about when she's in a conversation with developers and they're, you know, arguing about something, whatever it is. Like she will now understand at least on some level what's going on as opposed to just being like, well, I'm just marketing and I don't really know what they're talking about, right? like no now she knows what they're talking oh she's had that same bug she's had to where it didn't deploy she's had all those things right so she's had to think about security even if she's not an expert yet she's had to realize like where do I put the keys how does Laravel cloud help me secure my keys and things like that like now she knows all those things so very cool been very fun to uh yeah watch her over the weekend was very entertaining >> it was yeah I just I feel like a proud dad I'm just so happy it's [laughter] great >> and did this on her own accord, I assume. Like this is not this is not something like you've been pushing for just yet or have you been pushing people to build side projects or >> side projects? No. Um get used to building internal tooling and using agents to like help you do your job. Yes. >> Right. >> Um Okay. >> So, but you you know how it goes. You can't contain the bug. Once you get the bug, everything is possible. >> Everything's a nail. You got the hammer. There's there's a c there's a small chance that Sam's currently in the beginnings of her AI psychosis and I love that for her. I love that so much. Everybody has to do it. >> Another missionary out there for us. >> Yep. It's great. I love it so much. Um >> All right. >> I had a big I had a hit over the weekend. >> You did. You love a hit. There's nothing you love better. >> Oh man, I love a hit. I love a hit. So I think we're at like a quarter million views now. >> Oh wow. Okay. >> And the hit was I shared uh two screenshots of a prompt that I use uh to audit my codebase >> and then I shared the gist down below. Um and so people can just go grab it. But the uh the idea is I have had a like a very well scoped discrete prompt for a long time that I just call DSA and I use Solo's prompt templates and so I it's not like a skill or anything. I just inject the prompt myself. The DSA prompt is basically look back over the work we've just done. Cuz as as we've talked about many times, building is always this tension between uh figuring out what you want along the way and then once you know what you want, like cleaning it up. Your solution is Laravel new, right? Honestly, not honestly not a bad solution. I figured out I figured out what this product or what this code wants to be. Now let me redo it with a like an actual clear goal because even preAI part of building is discovering. I think a big part of building is discovering. So, I've had this DSA prompt for a long time >> that is it stands for data structures and algorithms >> and basically it says like look back over the work that you've completed and find uh a material simplification and here are the six eight areas you should look for scattered booleans that represent the same logical concept >> a shared assumption about like the shape of an object that we should just make into a proper first class object, right? >> Um, and so it's all of those like data structure things that don't they don't fix bugs, they eliminate the possibility that they could exist, right? So if you use an enum, you don't have these magic strings floating around and you just really don't have to like go back and audit are all the strings up to date? No, you're just you're using the enum. There's just no chance for it to drift. >> So that's been my discrete prompt for a while. I added on a layer um which is what I tweeted of auditing the whole codebase instead of just recently. Usually I do the DSA at the end of like a session. I'm like great we landed the feature look back and clean it up. >> This one is >> uh you basically have a an orchestrator catalog the entire uh universe of subsystems in your application, right? whether that's billing, web hook handling, whatever, right? Just go through and tell me what all the subsystems are. >> Then, and the reason for that is you don't want to send out an agent that's like, "Go make my code better." You're like, "I don't even know where to start, man." Um, >> so once you get all the subsystems cataloged, then you have the orchestrator send out a fresh agent into each subsystem with that DSA like prompt. So, inside of this billing module, module being a loose term, inside of this billing module, go look for material simplifications about boolean explosions and all of this stuff. >> And then it like brings it all back and then the orchestrator is supposed to catalog, ddup, rank, sort, all that kind of stuff. And it is incredibly useful. It is very good. So, I ran it on solo and it came back with I think it came back with almost a hundred suggestions and all of the suggestions are not like it's not like oh this might help. It's like no, you're already pretending that this is an enum and it's not. >> Just make it an enum and these things go away. And so that one uh that one has I think that one has hit a nerve because >> uh it's an actually like useful tangible audit instead of just like what would we do to make this cleaner? It's like no no no no no. You're pretending this is a state machine and it's not a state machine. Turn it into a state machine. >> So pretty happy with that one. >> So this is like what's interesting to me about this I know you had that cleanup one for a while. It never felt like the right level for me or something. Like I'm sure it's fine, but it's like after every feature doing it felt like maybe the AI gets a little like excited to go off and like find yeah find such things for you. You know what I mean? Whereas like so this one being more like hey we've done a bunch of stuff and I think I like for my use case when I'm doing a hubspot I'm going to use this probably with a slight tweak but it's like look at like all this stuff in this new migration project right we've done a bunch of stuff now like go through that and do this what your prompt says which is yeah go through it all look see the big picture now we've done a lot of work so there should be there probably duplicative things probably all yeah missing enumes or whatever and so it's got a lot to work with that's all still relevant to I just feel like it's got a lot to work with that is makes sense to look at at together, right? And then like to cut across the whole thing. Oh yeah, we should be doing this or that and then have it go through that and uh then yeah, and then find those patterns and whatever else we should be using. Um yeah, so I'm pretty excited to do that. I did something similar with a cleanup operation as part of this project last night where it's like uh basically there was like some steps in the process. I'm not going to get into the whole thing where like they're actually not that useful. So like let's kill those steps. Let's change all the rules I built cuz I have all this structure to it. We don't do those stupid steps anymore. We actually don't need them. Whatever blah blah blah. So yes, as you get deeper in, you see what you need, what you don't need. And this fits into that. I like I like this kind of thing. It's like, hey, one night just kick this thing off, let it come up with a bunch of and then uh I How do you handle Okay, so it makes a hundred suggestions. What's your step then? So my step then because I'm working inside of Solo, my step is the orchestrator is in charge of the gathering of the discrete agents findings and putting them into a scratch pad. Then when all of the exploration agents are done, the orchestrator has to go back through the scratch pad and basically like audit the audit, categorize them, group them, give them a P 0 score, P1, whatever. Um, >> and after that, so that is that's the final artifact from the overnight run. And I'm actually I'm leaning a lot more into overnight runs now that I have a job. Um, and I'm getting like I'm getting a sense of what's good for an overnight run and what's not. This sort of like super long audit. Fantastic. Because no code was changed. >> It spent, you know, half a billion tokens or whatever it was and I wake up to a hundred things that could be better. So I had this huge scratch pad of findings which is like I found this thing and this file it should be this and here's what this would obiate. >> Mhm. >> Then um I have an agent turn that into a set of linked solo to-dos. So now I've got this whole dependency graph of to-dos and then I put a new orchestrator on it and I say hey here's the dependency graph of to-dos. So that takes it from, you know, 60,000 characters of scratchpad to uh, you know, 500 characters of discrete problem. So it's kind of like we're we're finding all the discrete problems, putting them in a big thing, and then turning them back into discrete to-dos that then can then be fanned back out. >> Um, and then at that point, the problems are very well scoped and like very like explicit. And so I just have a I have a big dog on the orchestrator orchestrate, you know, soul high because I don't need extra high, I don't need max ultra. These things are already scoped. They should be relatively small cleanups. And so now I just have it on, in fact, it's on another computer. It's not even on my my MacBook. It's just like going through one by one, just cleaning it up, checking the tests, marking off the to-do, and moving on to the next one. >> Do you accept them all? Are you just assuming they're all legit, or do you read through it? Um, I have read through I've read through it. I didn't I didn't reject any. The orchestrator rejected several. There is some negative language in the prompt that's like don't just move stuff around to put it behind different boundaries. >> Um, and I've used the discrete DSA prompt so many times and every time that one comes back, I'm actively watching the terminal and I have to approve is this a good idea or a bad idea? And I've yet to see one that was offered to me that I thought was a bad idea on the discrete DSA props because oftentimes it'll come back and say, I actually don't recommend doing anything here. It would just be shuffling code around. And I'm like, great, commit it. Move on. We're done. >> Yeah. >> So, I have gained a lot of personal confidence in the discrete prompt. And that's what all of the exploration agents were given was the discreet prompt. And so, I feel pretty good about it. So, I didn't I didn't even read through them all. >> Yeah. Yeah. All right. Interesting. I'm excited to try this out. But yes, it's taking off. It's at 290,000. >> 290 >> 4 4,000 people bookmarked it. Who is bookmarks? I don't know. >> Yeah. Come on people. If you're going to bookmark, give me the like. Also, I think I have like one to four or maybe 1:2 one to four ratio on likes and bookmarks. Just like throw me a bone here. >> But not liking. And nobody retweets anymore, which I hate. Like I want the retweet. The retweet is the best part. It's like what made Twitter. The retweet is the best thing on Twitter. Nobody >> love the retweet. I did get the Addi Osami retweet who is like 15 to 20 year Google deep mind veteran who just recently left and he is like one of the most he's like he's the Jason Cohen clarity of thinker but for the AI like space he is just such a clear thinker. So when he retweeted it I was like ah okay this is good like I'm on to something here. Yeah [laughter] >> stuff's happening. We're going in the right direction here. Um, yeah, this is kind of funny. This is like I mean this is it's you know eight months ago we would have said, "Hey, this is like the kind of stuff you'd be releasing all the time as part of your faster stuff." So that's very interesting. You're still out there. You're still sharing with the people. Yes. >> Still in the game. >> Um, and now you have a bunch of stuff built up. You could be dribbling these out once a month. >> I know. >> Yeah. >> Um, all right. So big tweet. That was cool. I'm going to use it. I will report back how I do with it this week. I Yeah. Might might even Well, tonight we'll be busy, I guess. But the next couple days I want to do it because I'm kind of at an inflection point before I take on some of the really hard stuff. Uh, you know, I've been hitting the lowhanging fruit. So, kind of in this like reset phase of like, all right, let's double back, get rid of the stuff that's stupid. I'll run this thing, clean up architecture things, correct? and then uh you know handle the really heavyhitting pages that are going to be >> I do think that's a valuable pattern which is like make sure your make sure your foundation is pretty solid before moving on to the next giant thing you're going to layer on top of it. >> So I'm I'm currently building up um uh well I I will reveal it here for the first time. I'm working on the um the syncing of to-dos, scratch pads, and prompt templates um across your various machines. So, like if you have Solo on >> a MacBook and a Mac Mini or whatever. So, some people have a desktop and a laptop. >> Um it's frustrating that your scratch pads don't sync across those things. >> Yeah. And so something that I'm cooking up, I started uh I think maybe Friday night or Saturday um is and I've had this plan for a while is the um ability to sync those things. And interestingly, it is a content blind uh sync. So I never see it. I never see the content. I have no idea what it is. It's just this opaque set of bites that flows through a durable object. [screaming] >> You love durable object. love a durable object. So the durable object becomes the centralized store for this sinking mechanism. Uh and everybody gets a durable object. You get a durable object. You get a durable object. And so as I'm laying the foundation for this, I'm like before I move on to the next thing, I'm running this full DSA audit to clean up the foundation. And that part has been fun because this is the first time, you know, this whole DSA audit is very new to me. uh the orchestration layer. And so I've landed the foundation, cleaned it up, and now I'm moving on to the next part. And I'm like, I have this makes me feel really good. I'm I'm happy with this. >> Oh, solo. It lives on. So what's our solo update? Give us a little solo update. Like what you've been How's the time working out? How you feel about it? You know, all that stuff. >> Um I feel really good about it, actually. The time is working out. The time is working out fine. Um, I think some part of that is emotional where it's like I'm not I'm not my back's not against the wall on I got to get this stuff like >> it's a side project. >> It's a side project and it's taken off a lot of pressure which has made me a lot happier which has made me move faster. Um, [laughter] funny how that works. So, um, it's then this release is a little long in the tooth. Um, it's been maybe a month and I just keep like fixing things and making things better and so I have to stop doing that. Um, >> it's been a busy month though. >> It's been a busy month. Uh, but where we're at now is several like several quality of life improvements and hopefully several um adoption or like persist or uh like churn reduction improvements. So the big thing is uh workspaces >> which >> you know came with a lot more than I expected because solo has always been a single window application. >> Mhm. And the introduction of workspaces means you can pop out multiple solo windows which means you got to think about the routing of backend events and you got to think about state across these two web views now to make sure that when you read a notification on one side it like goes away on the other side. So there's been a lot of that and I'm really happy with where it's at now. I could not live without workspaces. So sorry to everyone that doesn't have them cuz I do and they're awesome. >> Yeah. I've also cleaned up a lot of the discoverability around uh the pallet. So like I use the command pallet religiously, >> but I also use all of the other pallets like command E is the quick jump, command P is like the quick actions, command shiftP is the prompt templates, and people are like I didn't I didn't I didn't know any of that existed. I had no freaking know any existed and I don't use any of it. Okay. >> So now, uh, command K is much more universal. Command K used to be just actions, but now it's like this amalgamation of everything. So if you hit command K, you should be good. But then also, when you hit command K, you see all of the other pallet hotkeys as sort of like a kind of like a tab bar. So if you're like, I'm going to hit command K cuz the UI says command K for actions, which it does at the bottom, you know, when you first use Solo. uh it leaves a little hint there until you hit command K five times. Then that hint goes away. But now when you hit command K, it's a natural entry point and discoverability point to the other pallets as well. So every time you hit command K, you see command E as your jump palette. You see command shift E as your like notifications pallet. So if something rings a bell, you can just hit command shift E and only the terminals that need your attention will show up there. Um, it shows the prompt templates, it shows the new tab stuff. So, I'm working a lot on making the whole thing easier to use and more discoverable. And then by far the best quality of life thing in the world is I've started capturing agent sessions immediately at startup. So like if you start a new codeex session >> and you type any any message immediately after that session gets its uid. So every one of these coding agents like they store their sessions by a UUID which is how you're able to resume a session. >> Now Solo is watching and when it figures out what the UYU it is, it immediately captures it and saves it into your database and shows a little like a little you're safe icon on the terminal like a little bookmark with a check on it which means you can quit solo hard. Like agents could be running and you could just quit it hard and then launch solo again and every every agent you can go back to and you have the option to resume that agent. Um instead of just being like, "All right, well, I'm going to type codeex- resume and look for my look for my session." No, they're all saved. Every single one of them, you can just go and hit resume and you're back to work. So that's that's by far my favorite quality of life improvement. >> Yeah, especially when you got >> a bunch going across a bunch of different projects. It's like yeah, it is annoying when you restart and then it goes south, which I have to restart a lot because my stupid Mac OS is freaking terrible. And so yes, that is super annoying. So I will use that a bunch. >> It already has kind of something like that, right? But it doesn't it wasn't reliable. It's you had to control C to quit your agent so that it would output onto the visible screen the uid which we would capture. So if you're listening now we can capture it already if you hit control C. >> Um but the improvement is >> we will capture it based on file system activity while you're still running it. So that if you do like if your machine does auto restart or you do want to just hard quit and walk away, you can do that without having to control C through them all. >> Yeah, cuz you have 20 of those open. It's like have to control or whatever. Yeah. So all right. Uh when's this coming then? When are we getting this? >> I think we're good. Also, I'll probably release it to beta the beta group tomorrow. Um, I'm doing one last pallet modification cuz something wasn't getting the right priority and it was a little confusing. Um, but I think I think we're good. I mean, it's it's made it through CI several times now. Um, >> I've done a lot of work on fixing all the flakes in CI because that just drives me freaking insane to have to push something and wait 30 minutes and it's like, oh well, this timeout wasn't set right for CI. I'm like, god dang it. Um, so I think tomorrow we'll get it out to the beta group of which you are a part, so you should be able to see it. >> All right, cool. All right, so you are going to Denmark tomorrow night or something like that. >> Going to Denmark. I leave tomorrow night. I am going to Laravel Live DK. Um, >> I am speaking. So I'm doing the same talk I did at Laravel Live UK. Um, which is what the hell are we supposed to do now? >> No, that's good. It's nice to do the same talk. It is. Um, and so that'll be a lot of fun. Taylor and Tom are going. Sam Sappenfield's going. A lot of a lot of Laravel people are going. Also, DHH is going to be there speaking. >> Look at that. >> So, he and Taylor are going to do a fireside chat. So, I think it's going to be a blast. I'm I'm very excited. Um, don't know much about Denmark, but uh I think I've been there maybe, and I'm led to believe it's awesome. So, >> are you going to have time to tour around or No, >> of course not. No. >> When you come back Saturday? Saturday. No, >> you got to tour around, man. We gota we gotta get you a little more time. These all these kids, they're they're young and they're little. >> Yeah, that's the only thing. I couldn't do that to Jennifer. I couldn't go on a fun vacation every quarter and be like, "See you." >> I have this question for you. Now that you are the vice president of marketing and community, what's your plan on these uh all the Laracons? Like, are you going to try to go to all of them? Some of them? >> Are you going to the big ones, not the little ones? Like, what are you thinking? I'm curious where you're at. >> I try to make myself available for every event that could that could use me. Um, >> of course, sorry, Michael, Australia, Australia is still very far away, >> and it's quite the tax on on my family. So, I I'm not going to make a commitment to go to all of them. Yeah. >> But I am trying I am trying to make myself available for as many events as would be helpful. Um, while while keeping that family tax in mind. Yeah. Okay. Well, I mean, the D deck here is an easy hop over the pond, so that's not too bad. >> And it always helps that it's during the week because >> there's school and there's the nanny during the week. >> And so that's why I keep it so tight. Like, would it be fun to stay through the weekend? Of course. But the weekends are the hardest for Jennifer because there's no school and there's no nanny. And so just really trying to turn around and get home at least for part of the weekend, if not all of it, is is my number one priority. >> Well, that makes sense. Soon soon they'll all be in school and they'll be relaxing and uh the weekends will still be bad, but you know, >> there's an argument to be made. I should be leaving tonight so I get an extra weekday in Denmark instead of staying an extra day at the end. >> Oh, yeah. >> Which I didn't think about until this very moment. Switch the flight. Hop on a flight tonight. Get out there. >> No way. >> Uh, we got a podcast to do tonight. >> Tonight. >> Um, all right. I don't know. Anything else? >> Tell tell the people what's going to hit their feeds. We've we've mentioned it a couple of times. So, you're going to hear this one and then what are they going to hear? They're going to hear >> Sweet Hang with Jesse Hanley. We're going to fight about AI. We are going to um talk about camera gear. So, you'll be excited about that one. And I don't know, I have a couple other things on the list, but we got we got some topics lined up. Um, we've been going back and forth with him a lot on Twitter and seemed like we should we should hash it all out on the podcast. So, I think that was the main idea. Um, he's been doing a lot of cool like video marketing sort of marketing kind of things that I want to talk about, too. I got to put that on the list cuz those seem to have been very effective. Um, >> the whole founder to the people thing is working really well. >> Yeah. And people love the video. I got to do more video. This is like a thing I want to do with uh some probably HubSpot stuff and that was always already the plan with outro if it ever releases. We will see. >> Um but yes, so that I think is what's hitting the feed. That's what I don't KNOW WHEN WHAT I DON'T KNOW. We are Do we want to save these for like Thursday? Do we want to just roll them out as soon as Dave producer Dave's got them ready? What do you think? >> I'm I'm open to either. I my hottest take in the world is that there aren't enough podcasts. So, I'm always I'm always like, "Let her rip, baby." Uh, but I don't I don't >> That's my inclination. I guess Yeah, I guess we'll see. All right. Maybe we'll do Tuesday and Wednesday or Tuesday and Thursday. >> Especially cuz our last double episode was like very of the moment. We were talking about stuff. That's why I was like, let's get this out the door. This Jesse one is more evergreen and so we could we could push it to Thursday. But let us know what you think in the comments. If you're just yearning for more content, just tell us and we might try to ship it early for you. >> Yes. If we get a bunch of comments Tuesday morning about how we must have Jesse Hanley Tuesday night, we can make anything happen. It's all possible. >> We sure can. Through the power of Dave, [laughter] >> we're not doing a thing. >> Yeah, we can't do anything. We can't trust us to do anything. It's not It won't happen. >> We're exec producers. What do you want from us? >> Oh, we should be listed as not just hosts, but also executive producers. Then I got my executive producer title. When are you gonna get me into Laravel as some kind of executive producer? I want to be an executive producer. >> I know. I know. We don't we don't have a lot of sit around and do nothing and still get [laughter] paid jobs. But if any of those open up, I will let you know. >> You know what I want? I don't really want a job at Larel. But if I was going to have a job at Laravel, I want to be on the board cuz that's like almost sit around and do nothing. But like ta Taylor needs a board member in his corner. You know, he got all his VC board members. >> Need a Taylor board member. >> Yeah. me. So, when the board expands, >> you would be a great board member. >> Who would be a better board member than me? >> I don't I actually don't know. You'd be a great member. >> So, I'm just putting it out there to the universe when the board expands. >> Well, we'll talk we'll talk to uh President Tom and see if we can get all Ian on the board today. >> Tom's already on the board. We need on the board. >> That's what I'm saying. >> Yes, >> he's the president of the board. So, >> put it out there. Yeah, that's a good idea in the ether. >> You'd be a great board member. >> Yes, I know. >> Anyway, all right. Thanks for listening. Thank you to the sponsors. Check us out at mostlytechnical.com for the old episodes and new episodes and links to everything. And most importantly to sponsor the show. Links are there. Um yeah, we will see you actually in a day or two with another episode. And then we'll see you next week and Aaron will tell us all the shenanigans he got up to in Denmark, which should be exciting. So, we will uh talk to you later. Thanks., 145: Marketing as Engineering, Ian Landsman & Aaron Francis (Mostly Technical), 01m0ar614fzmvzgv62399k4av3.