When the Ralph autonomous loop was born, Dex Horthy was "in the garden," witnessing the spark that set the AI engineering community on fire. Andrew sits down with the HumanLayer founder to discuss how to escape the "Dumb Zone" by applying his strict RPI (Research, Plan, Implement) methodology - a process that forces agents to generate intermediate design artifacts and align on architectural decisions before writing a single line of code. They also break down the brutal economics of agentic coding, recounting how Dex’s team used autonomous loops to clone six sponsor products overnight at a hackathon.
Show Notes
- No Vibes Allowed: Solving Hard Problems in Complex Codebases: Dex’s viral talk
- The AI Vampire: Steve Yegge’s reflection on burnout and agentic economics.
- 12-Factor Agents: HumanLayer’s methodology for building reliable agentic systems.
- Twitter: @dexhorthy
- LinkedIn: Dexter Horthy
Transcript
(Disclaimer: may contain unintentionally confusing, inaccurate and/or amusing transcription errors)
[00:00:00] Andrew Zigler: Welcome back to Dev Interrupted. Earlier this year we sat down with Geoff Huntely for a conversation that frankly set our community on fire.
[00:00:08] Andrew Zigler: He introduced us to Ralph the autonomous bash loop that And he's like, lemme also go put in a good work with you through this guy called Mike Web.
[00:00:15] Andrew Zigler: that the
[00:00:16] Andrew Zigler: like, because I,
[00:00:17] Andrew Zigler: had just been
[00:00:18] Andrew Zigler: okay.
[00:00:19] Andrew Zigler: he
[00:00:19] Andrew Zigler: Yeah. 'cause apparently they do
[00:00:20] Andrew Zigler: throughout that
[00:00:21] Andrew Zigler: together.
[00:00:22] Andrew Zigler: one
[00:00:22] Andrew Zigler: I just.
[00:00:23] Andrew Zigler: up as a catalyst for many of these discoveries. You know, the other person in the room, the, he was in the garden kind of energy and. Dex Horthy and Dex was there in San Francisco when Ra, the Ralph Meme was born. You know,
[00:00:36] Andrew Zigler: Okay, cool.
[00:00:37] Andrew Zigler: the brutal unit economics of the new
[00:00:39] Andrew Zigler: Nice.
[00:00:40] Andrew Zigler: famously
[00:00:40] Andrew Zigler: great.
[00:00:40] Andrew Zigler: napkin.
[00:00:41] Andrew Zigler: And today we're bringing Dex on to pick up the story where Jeff left it off and dive
[00:00:46] Andrew Zigler: Sure.
[00:00:47] Andrew Zigler: happens when the loop becomes the standard. You take it to scale with an engineering team, and you're not making something new, but you're tackling those brownfield problems that are messy and are really hard to deal with, with agentic stuff.
[00:00:57] Andrew Zigler: So Dex, welcome to the show. We're so excited [00:01:00] to dig into this with you.
[00:01:01] Dex Horthy: I'm so excited. I read the list of questions you prepped and I'm like, oh, good. We was just gonna do story time. I'm, I'm stoked. I love, I love story time podcast. So, uh, I, we'll get into the details. We'll do some, some tech stuff, but yeah, I'm, I'm excited to, uh, to, to riff with you and see where it takes us.
[00:01:18] Andrew Zigler: exactly. I know, I, I know, I know. You're typically, like you said on the whiteboard and breaking things
[00:01:22] Andrew Zigler: Doesn't, so I was trying to understand like if we, if we have the same contact
[00:01:26] Andrew Zigler: And
[00:01:26] Andrew Zigler: so
[00:01:27] Andrew Zigler: different
[00:01:27] Andrew Zigler: today we
[00:01:27] Andrew Zigler: person,
[00:01:28] Andrew Zigler: things a
[00:01:28] Andrew Zigler: uh, but it sounds like.
[00:01:29] Andrew Zigler: different speeds, more
[00:01:30] Andrew Zigler: Different person in this case.
[00:01:31] Andrew Zigler: we're gonna go on story time for sure. I, I
[00:01:34] Dex Horthy: Amazing.
[00:01:35] Andrew Zigler: by framing it. Um, about when you were with,
[00:01:38] Andrew Zigler: Hmm,
[00:01:39] Andrew Zigler: because I
[00:01:39] Andrew Zigler: interesting. Okay. Let me, let me hit up the guy then I know at liquidity
[00:01:43] Andrew Zigler: was
[00:01:44] Andrew Zigler: and tell him like
[00:01:45] Andrew Zigler: of the
[00:01:45] Andrew Zigler: to, or be like, be like, oh, I have another point of contact or
[00:01:49] Andrew Zigler: rocking out at a
[00:01:50] Andrew Zigler: through for this.
[00:01:51] Andrew Zigler: you know, ended up talking for a while.
[00:01:53] Andrew Zigler: What, what, what
[00:01:54] Andrew Zigler: Okay, cool.
[00:01:55] Andrew Zigler: and what, what, what were the events that kind of led to deeper conversations with Jeff
[00:01:59] Andrew Zigler: [00:02:00] Yeah, I mean,
[00:02:00] Andrew Zigler: building.
[00:02:02] Dex Horthy: yeah, so, uh, around, I think I wanna say like February or March of last year, a buddy of mine who I'd worked with a long time ago, dropped me in this like random Twitter group DM with about 50 people in it. Uh, and it was, um, basically people riffing about models, about running their own models, about getting their GPUs to turn out more tokens per second complaining.
[00:02:28] Dex Horthy: It was like before Kim EK two had come out, but people were running open source models. It was. Not a lot of Claude code until one day it was all about Claude Code. Um, but it was just like agentic coding plus more generic AI stuff. Um, and the person who runs that group chat, she put together basically like a small meetup for like the people who are live in sf.
[00:02:50] Dex Horthy: Um, and we hosted maybe like 10, 12 people in person in our office. And then another like five or six people like joined over the Google meet and we just did like five minute lightning talk show and [00:03:00] tell of like. What's your favorite? A agentic coding thing? I think this was in like June of 2025
[00:03:06] Andrew Zigler: Wow.
[00:03:07] Dex Horthy: and it was the first time I saw a bunch of really cool tools that I actually haven't heard of in a while.
[00:03:11] Dex Horthy: It's like some of the things are like, yes, you still hear about that other day every day, and some of them were like, I haven't heard of, so things like Spec Story and Task Master, which were some of the early like CPS that people were using to do this kind of like thing that's now very in vogue, which is like the beads or agent swarms thing where you give the the coating CLIA tool to be able to like manage work.
[00:03:30] Dex Horthy: You know what I'm talking about? Mm-hmm.
[00:03:32] Andrew Zigler: Oh, I'm, I use beads. Yeah. It's
[00:03:34] Dex Horthy: Okay.
[00:03:34] Andrew Zigler: now. I know what you're talking about.
[00:03:36] Dex Horthy: Yeah, so there were like early things that were in the beads direction that were, uh, that never really caught on the way the beads did. Uh, and so that, that was super interesting to see, um, kind of those kind of things. There was people, it was the first time I saw context seven, which was this tool that would let, it's like now very popular.
[00:03:54] Dex Horthy: Super. Yeah. Yeah, and since then we've been using it. We're huge fans of Contact seven. So it's just like people [00:04:00] sharing tips and tricks of like, here's this random thing that no one's heard of yet that we're playing with. And um, my buddy who had originally like introduced me to this crew, he ended up showing up like two and a half hours late and he brought with him this guy Jeff, that no one had ever heard of.
[00:04:15] Dex Horthy: Uh, and uh, he came in and he was sitting and like took in the last couple talks and then we were like, Jeff, do you want to go and like show us your thing? And he just launched into this like 30 minute, like deep dive into Ralph and how it worked and showing us the live streams of like, yeah, literally I'll turn this thing on, I'll go to bed in Australia, I'll leave the live stream on for 12 hours and then I'll wake up and like check on the code.
[00:04:39] Dex Horthy: That's why I also thought liquidity would be an interesting, uh, idea because,
[00:04:42] Andrew Zigler: and
[00:04:43] Dex Horthy: um,
[00:04:43] Andrew Zigler: some of his
[00:04:44] Dex Horthy: they're big.
[00:04:44] Andrew Zigler: 'cause he had been building like the cursor standard, uh,
[00:04:47] Dex Horthy: Um.
[00:04:48] Andrew Zigler: terms of like all these very structured rules and how he was building his system. And we were learning a lot from him here.
[00:04:53] Andrew Zigler: And so, uh, I'm totally familiar with the whole, just like, okay, I'm in a
[00:04:57] Dex Horthy: Yeah, that'd be great. That'd be great.
[00:04:58] Andrew Zigler: this whole
[00:04:58] Dex Horthy: Yeah, I think that's probably the best [00:05:00] move.
[00:05:00] Andrew Zigler: and then ultimately kind of a, you know. Belittle
[00:05:02] Dex Horthy: Oh yeah.
[00:05:03] Andrew Zigler: it became pretty popular
[00:05:05] Dex Horthy: This sounds great though. Thanks.
[00:05:06] Andrew Zigler: terms
[00:05:06] Dex Horthy: Thanks. Thanks, Evelyn. Um, I don't need to like,
[00:05:08] Andrew Zigler: a way to
[00:05:09] Dex Horthy: um,
[00:05:10] Andrew Zigler: that
[00:05:10] Dex Horthy: sound like I'm trying to like
[00:05:11] Andrew Zigler: and then
[00:05:12] Dex Horthy: step over you or anything. It is like, the truth is like these deals are, are really tough.
[00:05:17] Andrew Zigler: your
[00:05:17] Dex Horthy: Yeah.
[00:05:17] Dex Horthy: Yeah. So it's like if we're gonna do it.
[00:05:19] Andrew Zigler: real. It's kind of how these things happen.
[00:05:21] Dex Horthy: I appreciate that.
[00:05:22] Andrew Zigler: group
[00:05:22] Dex Horthy: I appreciate that.
[00:05:23] Andrew Zigler: with a bunch of other wacky people experimenting with things.
[00:05:26] Dex Horthy: Amazing.
[00:05:27] Andrew Zigler: some
[00:05:27] Dex Horthy: That's great to hear.
[00:05:28] Andrew Zigler: and
[00:05:28] Dex Horthy: Okay, cool. Go outta your video and then send it to me. Yep.
[00:05:31] Andrew Zigler: are
[00:05:31] Dex Horthy: Yeah.
[00:05:31] Andrew Zigler: out these primitives and how they scale and how to build with them.
[00:05:35] Andrew Zigler: And it's not that it's like, uh, un unsolvable. It's just that you have to be willing to get in there and get messy, you know?
[00:05:40] Dex Horthy: Well, and it's, I think it's really hard to create those sorts of communities. Like I've been in SF for 15 months now, and the thing that I said, like we were building AI Dev tools since like. Summer of 2024 and it was like, oh my God, there's seven meetups a week. Like you can just literally go talk to users.
[00:05:56] Dex Horthy: They're everywhere compared to, I was in Chicago before and like there's some [00:06:00] good AI engineers there, but they're hard to find and they don't get together that often. And then within like a couple months, I was like very burned out on the SF meetup scene because it was just like literally every event is the same.
[00:06:11] Dex Horthy: You show up, you get some mediocre pizza. You hear some, you hear six talks. Three of them are like lightly veiled vendor pitches. Three of them are like full on vendor pitches. And so it was just like, I just noticed that like the people that were actually at the frontier were finding their own spaces to get together and a lot of it was happening on like Twitter and like small discord groups.
[00:06:32] Dex Horthy: And so like, I don't know, I do a lot of organizing in SF now, and I think a lot about like how can we create. Space that doesn't feel over commercialized or like you are the product. Uh, and like
[00:06:44] Andrew Zigler: space. Like that's what
[00:06:45] Dex Horthy: Exactly.
[00:06:47] Andrew Zigler: Right?
[00:06:47] Dex Horthy: Yeah.
[00:06:48] Andrew Zigler: pitch zone. I know exactly what you're talking about. Like there was a period of time for sure. Where it was like, God, like another MCP night, another MCP night, another MCP, and you know, like those kinds of
[00:06:57] Dex Horthy: Yeah.
[00:06:57] Andrew Zigler: so it's like, but that's not really where people were [00:07:00] discovering these new kind of primitives.
[00:07:01] Andrew Zigler: That's not where Ralph Loops were getting born. That's not where they were getting shared either. And so, um, of like talking about, Jeff comes in this new character and shares this. This, this, this way of building with Ralph. And then afterwards y'all are talking more. And, um, from what I understand, like in terms of how he put it, that, know, y'all started breaking down the economics of like, if you had a Ralph Loop bill for you, then what does that mean for the cost and the, the value of a software developer?
[00:07:28] Andrew Zigler: Like what are your thoughts on that? I think the number he gave it was 10 42 an hour for a Ralph Loop to do software engineering work.
[00:07:35] Dex Horthy: Something like that. Yeah. So like, uh, a couple months later I was talking to, it was like August, so like two months later I was talking to a buddy of mine who had read the Ralph stuff but had never tried it. And he said like, let's go do this hackathon of Y Combinator. Uh, and I wanna do something about Ralph.
[00:07:49] Dex Horthy: Wanna learn to use it. I wanna build something cool. We sat around riffing about what we could do and it was like, okay, could our Hackday project be like a tool that helps you create Ralph Loops or set them up? And then we had this idea of like, what [00:08:00] if we just use Ralph? 'cause when you do a hackathon, they want you to do thing again.
[00:08:03] Dex Horthy: They want you to do things related to the sponsor tools. They want you to, yeah, there's seven sponsors and the more tools you can put together the, the better score you get from the judges and stuff. Right.
[00:08:13] Andrew Zigler: Mm-hmm. Yeah.
[00:08:14] Dex Horthy: But we decided to take a note if we were gonna use Jeff's technique, we would also highlight maybe some of his, uh, perhaps, uh, well-intentioned irreverence.
[00:08:22] Dex Horthy: And so, uh, we decided to take all of the vendor projects that we could and spin up a bunch of Ralph Loops to clone them. So it was like one of the sponsors was browser, browser use, which is this Python library that's really good for browser automation. We're like, okay, what if we ported the whole thing to TypeScript?
[00:08:39] Dex Horthy: And so we just like literally set up a tool that would just run overnight and port all of that to TypeScript.
[00:08:45] Andrew Zigler: Yeah, so
[00:08:45] Dex Horthy: Um,
[00:08:46] Andrew Zigler: build tools that eat them.
[00:08:48] Dex Horthy: exactly. Yes, exactly. Um,
[00:08:50] Andrew Zigler: Yeah. Which I
[00:08:51] Dex Horthy: and I remember the,
[00:08:52] Andrew Zigler: now.
[00:08:53] Dex Horthy: there's a picture of the blog post of like the, the CEO of browser use looking at it as like, holy shoot, you guys did this overnight, like without [00:09:00] looking at it and it like, kind of, I mean, it wasn't perfect.
[00:09:02] Dex Horthy: I think the thing with the Ralph stuff that people, people say, oh, this doesn't work. Yeah, but it gets it like 90% of the way there and like it's much better than if you tried to directly vibe code the whole thing or something like this.
[00:09:14] Andrew Zigler: Yeah, the sticker, like the shock, the shock factor, like the eye getting big at the moment, like those have been real
[00:09:19] Dex Horthy: Yeah.
[00:09:20] Andrew Zigler: Like I, uh, like two weeks ago, I, I won an AI hackathon for The Atlantic, which is like, you know, the Journalism Institute. So you're exposing them to like, how you can
[00:09:29] Dex Horthy: And uh.
[00:09:30] Andrew Zigler: like, uh, you, how you could build new technology on top of their preexisting like archives and make them more accessible and digestible.
[00:09:37] Dex Horthy: And I was like,
[00:09:37] Andrew Zigler: what
[00:09:38] Dex Horthy: what did, what did you, what did you pitch them? That one,
[00:09:39] Andrew Zigler: was helping other people because like
[00:09:41] Dex Horthy: this is the worst part.
[00:09:42] Andrew Zigler: And then when I presented it with them, it
[00:09:44] Dex Horthy: You pitched them 10 K
[00:09:46] Andrew Zigler: because they were
[00:09:46] Dex Horthy: and they said That was high. And I was like, I was like, mm, mm-hmm.
[00:09:51] Andrew Zigler: was served.
[00:09:51] Andrew Zigler: Like they were con and then I left. I got an Uber and left and
[00:09:54] Dex Horthy: I told her like, don't worry. I.
[00:09:55] Andrew Zigler: it's like, you know, the mystery continues. And so I think it's funny to hear like, you [00:10:00] know, you're gonna
[00:10:00] Dex Horthy: Deal.
[00:10:01] Andrew Zigler: and you're gonna use as a, as a
[00:10:03] Dex Horthy: And I was like, if you and I and she said, can you send me your rating? Can you send me
[00:10:06] Andrew Zigler: how
[00:10:06] Dex Horthy: your rating? Because I don't want her to go.
[00:10:08] Andrew Zigler: Um,
[00:10:08] Dex Horthy: So I was like, look, if you think this guy is generally good deal. Yeah.
[00:10:11] Andrew Zigler: I I, I
[00:10:12] Dex Horthy: I was like, you can just directly,
[00:10:13] Andrew Zigler: that I read, I think just
[00:10:15] Dex Horthy: but otherwise I have another person I'm working deal.
[00:10:17] Andrew Zigler: uh, his most recent piece, the AI vampire, where he reflected, uh,
[00:10:21] Dex Horthy: Oh, I didn't read that one yet,
[00:10:23] Andrew Zigler: Oh yeah. So that one just came out yesterday. So
[00:10:25] Dex Horthy: girl.
[00:10:25] Andrew Zigler: in this one about, um,
[00:10:27] Dex Horthy: she making?
[00:10:28] Andrew Zigler: of like being a 10 x engineer, a hundred X engineer, and the insane value extraction pressure that happens when you're taking advantage of orchestration.
[00:10:36] Andrew Zigler: When you're
[00:10:36] Dex Horthy: Mm-hmm.
[00:10:37] Andrew Zigler: and you have all these benefits, what does it mean for you and your teammates who captures that value? How
[00:10:42] Dex Horthy: Yeah,
[00:10:43] Andrew Zigler: from
[00:10:43] Dex Horthy: it's really tricky. Um,
[00:10:45] Andrew Zigler: Self-reflective
[00:10:46] Dex Horthy: I don't know why people won't like sometimes feel like just,
[00:10:48] Andrew Zigler: times vocally about burnout,
[00:10:50] Dex Horthy: yeah.
[00:10:50] Andrew Zigler: many organizations.
[00:10:51] Andrew Zigler: So, um,
[00:10:52] Dex Horthy: Interesting.
[00:10:53] Andrew Zigler: I know you haven't read it, so I'm not gonna go press you on it, but I'm just curious what your thoughts are on like, um, how employees can
[00:10:59] Dex Horthy: Like, [00:11:00] the only time it makes sense is if like,
[00:11:01] Andrew Zigler: If
[00:11:02] Dex Horthy: like doing something
[00:11:03] Andrew Zigler: and like
[00:11:03] Dex Horthy: like, like
[00:11:04] Andrew Zigler: overnight that
[00:11:04] Andrew Zigler: like
[00:11:04] Dex Horthy: the deal, like the deal with,
[00:11:05] Andrew Zigler: alive, then
[00:11:06] Dex Horthy: um,
[00:11:07] Andrew Zigler: really
[00:11:07] Dex Horthy: like through
[00:11:08] Andrew Zigler: your
[00:11:08] Dex Horthy: was like smart, right? Yeah. I, um, I don't know. I mean, I also, I think, I mean one thing I, I think I said online months ago is like. If you're working, 'cause we developed this methodology as like RPI research plan implement, which is not like unique.
[00:11:26] Dex Horthy: Lots of people were doing this. We just put a lot of time into like tooling and, and kind of, uh, guardrails for doing it well, prompts and all of this. Uh, but it's really designed for like brownfield code bases and like, hey, you have an existing software thing that you can't just like throw out and rebuild.
[00:11:41] Dex Horthy: Ra, but people are like, oh, I'm building a new project. How do I use RPI for it? I was like, what are you gonna re There's nothing to research, there's no code. It's greenfield. So like, if you're doing Greenfield, just write the specs and use Ralph and maybe use Ralph to help you build the specs and things like, so like, I think a, a gas town or Ralph, like, looks really, really good for like new stuff I.[00:12:00]
[00:12:00] Dex Horthy: We have a couple, I'm happy to chat about a couple of, like the applications of Ralph that work for us. Building a software product that is, I mean, it's not Brownfield, it's, you know, six months old, but it's, you know, 50,000 lines of code and like it has to continue to work. We can't just, I mean, the thing with beads is like 250,000, 300,000 lines of code.
[00:12:18] Dex Horthy: It's great. It's rif. No one's ever looked at the code. That's fine. Because no one's like paying money to use it. And if it breaks, it's like, cool, it's open source, it's free. Like what do you want? Kind of thing versus production. Exactly.
[00:12:30] Andrew Zigler: Yeah.
[00:12:31] Dex Horthy: Um, but the way, coming back to your first question, which is like, how did we get to the like 10 42 an hour or whatever is like, we did this hackathon and we spun up basically like we spent until two in the morning.
[00:12:40] Dex Horthy: We're like setting up GCP VMs and like getting the credentials put in and like setting up the team M sessions and running these, each Ralph loop for each of the like six sponsor products to try to clone them.
[00:12:51] Andrew Zigler: Yeah.
[00:12:52] Dex Horthy: we, I think it was like something like six, $700 in, I mean we had, I didn't throw out my credits, so we just used my credits.
[00:12:58] Dex Horthy: Uh, [00:13:00] but we did the math. It was okay, six servers, six, seven or eight hours, 600. It came out to about 10 40 or $11 an hour or something to run sonnet in a loop forever.
[00:13:11] Andrew Zigler: That's incredible. So you go and you're like, okay, I'm gonna spec out the machine, I'm gonna get the tokens. I'm gonna set all up, and then I can just amortize it and be like, this is how much it's gonna cost to just print execution for me to create whatever this software idea is gonna be. And obviously it's like that's. One challenge when you're doing something for a hackathon or a greenfield kind of space, or just replicating a reporting a project, but you know, it actually, it actually bridges us a bit into the realities of using those, um, uh, types of patterns on pre-existing projects and actually
[00:13:39] Dex Horthy: Yeah.
[00:13:40] Andrew Zigler: as like a career engineer.
[00:13:41] Andrew Zigler: Someone who all, I own a product, I ship things that are used by, you know, thousands or hundreds of thousands of people that are they're system critical. Like I can't even risk. Things. So,
[00:13:51] Dex Horthy: Yeah, if I get paged at three in the morning,
[00:13:54] Andrew Zigler: Yeah.
[00:13:55] Dex Horthy: like if I get page at three in the morning, I'm not gonna just poke my Ralph loop a couple times. I [00:14:00] have to be able to get in the weeds and fix it, whether I'm using a coding agent or doing it by hand, like it has to be. I can't be like, yeah, I've never read this code.
[00:14:06] Dex Horthy: I don't know how it works.
[00:14:07] Andrew Zigler: Yeah, precisely. So, um, even, it even goes to a bit about like, um, I guess jumping from some of the origin of like, okay, you have Ralph, it allows people to kind of like loop this and, and we've explored Ralph here quite a bit and how the economics and the value changes. But then when we talk about actually applying it, this is where we get to some really interesting conversations about the applied engineering and the reality of.
[00:14:30] Andrew Zigler: Being a, using orchestrator types of patterns, but then also working with these, um, with these tools in preexisting code bases. Like, uh, something that you've really championed is the idea of the dumb zone. And I love the dumb zone. You, you recently gave a talk. It was, um, uh, at, uh, I think it was a IE AI engineer was that was at
[00:14:49] Dex Horthy: Oh yeah. In New York.
[00:14:51] Andrew Zigler: Yeah.
[00:14:51] Andrew Zigler: that that, that talk is amazing. And we've included in our roundup before, it's definitely gonna be accompanied this episode as well. Um,
[00:14:58] Dex Horthy: Amazing.
[00:14:59] Andrew Zigler: it very [00:15:00] religiously because your ability to break down, uh, the way that like context, like actually needs to get constructed and then utilize, and then how a lot of the patterns that we assume are like, oh, this will make me safe, are actually like. Wasteful and they don't scale. So you
[00:15:15] Dex Horthy: Well, that was.
[00:15:16] Andrew Zigler: you call, you call bear the real reality of it, or I'm like, I'm tired of cycling my, my propo, my specs and my proposals and my tests every single time I change my idea. And so it's like you clearly felt the same. Yeah.
[00:15:26] Dex Horthy: Yeah, I mean that, that was also like the, the, the beauty of Ralph when it came out is like, you know, I don't use it to build most software we have in our, in our IDE product, there is a Ralph inspired thing where you have like a parent agent that owns the implementation of, you know, a plan that might be up to a thousand or 1500 lines of code.
[00:15:44] Dex Horthy: And it like shells out the, the phases as sub-agents. And that's how we do context isolation. You have a dumb model, write the code, you have a smart model, kind of check it, and then rete and then. Launch the next one. But the, the beauty of Ralph was this idea underneath it, which is like [00:16:00] everyone's got their, I mean there was no gas town at that time, but everyone had their like multi-agent system and super like complicated orchestrators and all this stuff.
[00:16:08] Dex Horthy: And it was like, no, as long if you actually know how contact windows work, like the only thing that really matters is like, how do you optimize for staying in the smart zone? How do you optimize for like small in digestible tasks and resetting context all the time? Uh, and everything else was like way over complicating and actually like quite simple problem.
[00:16:28] Andrew Zigler: Yeah. It actually, I, it is a really great way to phrase it that way, because. Ultimately what Ralph showed us is that the complexity is actually something you wanna strip away and you wanna get as simple as possible. A, a lot of people were using AI and AI engineering as an excuse to glom on more complexity and to handle these like levels of extreme complexity that they hadn't weren't able to do before, but then they were ultimately creating these like crazy.
[00:16:51] Andrew Zigler: Bike projects that would just fall over the second that anyone would look at them weird. And so like that's, and it's like how can you, that's not actually the benefit that you [00:17:00] gain long term from it. The benefit you gain from long term from it is how do you get smaller? How do I make it to where it's fail proof to where this tiny little loop. Can't get it wrong.
[00:17:09] Dex Horthy: Yeah.
[00:17:09] Andrew Zigler: what I, I mean, that's what I've loved about like cooking things with beads is because if I can make a whole bunch of little tiny atomic beads and then I can shove it to, like you said, a dumb agent, like someone who doesn't need to know anything, that can just
[00:17:20] Dex Horthy: Yep.
[00:17:20] Andrew Zigler: something very clearly, um, you start to get this division of roles, and this is where you get like the orchestrator patterns, where you have the smart agent, you have the, the smart human operating, the smart agent to vary. Carefully construct these contexts, you know, rivers that then these like downstream, like very dumb agents that are naive, but just really good at triggering stuff can then execute on. So like, uh, what, where do you like, see uh, that kind of thing going? Uh, the dumb zone is obviously an optimization that you can solve for now.
[00:17:54] Andrew Zigler: Uh, do you think that's gonna continue to be a primitive, or do you think that there's something that goes beyond that?
[00:17:59] Dex Horthy: Um, [00:18:00] great. That's a great question. Yeah, I mean, we talked about this in the 12 factor agents talk back in June and this comes up, I literally gave a lunch and learn yesterday. Um, and the question was like, well, when the models get smarter, do we still have to worry about all this stuff? I mean, the answer is like, as the models get smart.
[00:18:15] Dex Horthy: Okay, so here's, here was my experience in December. All of the good engineers who were kind of anti ai, like OG infrastructure. I mean, Mitchell Hashimoto's, the exception, he's been doing AMP stuff for six months or whatever. But a lot of engineers who were kind of skeptical of AI suddenly came around like, Opus 4.5 was the turning point.
[00:18:32] Dex Horthy: And I'm like,
[00:18:32] Andrew Zigler: turning point. Yeah.
[00:18:33] Dex Horthy: guys,
[00:18:34] Andrew Zigler: people and everyone would wanna winter break and everyone came
[00:18:36] Dex Horthy: yeah.
[00:18:37] Andrew Zigler: and now GitHub can't even stay online.
[00:18:40] Dex Horthy: So maybe they're doing too much Ralphing. Uh,
[00:18:43] Andrew Zigler: I.
[00:18:43] Dex Horthy: so the thing that, the, the thing that I think happened is like. It became possible to get the, the model got smarter and so you could get good results without doing as much context engineering, without having as much inex intuition for models, without really knowing how all this stuff works.
[00:18:59] Andrew Zigler: Yeah.[00:19:00]
[00:19:00] Dex Horthy: Um, or without like basically context, maxing, smart zone, maxing, whatever you wanna call it. Those people are now getting the results that engineers, the best engineers. I knew people like Jeff and many other people were able to get last summer with Opus four. And so it's like, imagine what those people are able to do now with Opus five, op Opus 4.5, Opus 4.6 if buy up.
[00:19:21] Dex Horthy: So it's like the models are gonna keep getting smarter and the way that I think about it is really like the, the notebook LLM team described this really, really well. They built a great product, um, and the way they did that, it was like they found a thing that is. J, the only way to build great experiences in AI is to find a thing that is like right on the boundary of what the model is capable of.
[00:19:40] Dex Horthy: It gets it right some of the time and then you find a way to context engine near your way into getting it right consistently. And so even as that frontier expands, as the models get smarter, the hardest thing they can do and the thing they can only do reliably, if you really think about it and are tasteful about it, it gets bigger.
[00:19:57] Dex Horthy: But imagine what the people who were rocking [00:20:00] opus for and getting crazy good results are getting now with this even smarter models when they're willing to do this context engineering and frequent compaction and stuff like this.
[00:20:08] Andrew Zigler: Absolutely. It goes back to like the star shaped intelligence. We've all seen like the Venn diagram where it's like the human intelligence is the circle. Then you got like the AI's intelligence, which is like all these spikes coming off
[00:20:17] Dex Horthy: Mm-hmm.
[00:20:18] Andrew Zigler: those same kind of spikes can be built and then reinforced with context engineering, which is in fact like what our jobs become as engineers.
[00:20:25] Andrew Zigler: And I think that's what I love most about, like when, when Jeff came on the show talking about like, well, you know. You're an engineer, aren't you? Like engineer the problems away if like you're
[00:20:34] Dex Horthy: Yep.
[00:20:34] Andrew Zigler: these, these systems capture the back pressure, find ways to turn the problems into something that drives the solution, and that really resets the conversation and brings engineers back to the table. Um, and so I, I just like love the way that he had, he had framed that. But I, I also wanna ask too, just because, you know, we've talked a little bit talking to other people. but I wanna talk a little bit more about you, d And so you know, what you are working on at Human Layer I think is [00:21:00] largely also tackling this issue of ha of, also goes back to like context engineering and having the, the information that you need to work on, like things that, you know, code existed that that existed before AI came on the scene.
[00:21:12] Andrew Zigler: How do we keep shifting that code? But with ai, so like, what are like the, what are the problems that go through your head with all of. This fast moving space, you're moving in and then like how do you apply it into what you're building now? What, what opportunities do you see as a founder?
[00:21:26] Dex Horthy: Yeah. Um, so I think the way I've been framing it recently, um, and we've evolved this stuff. It's funny, I got, I got on stage in, uh, in, uh, in November and also in August when this fir the talk first, the, the talk that was like the precursor to the AI engineer talk. And I guy was like, well, you have this thing, it's called RPI.
[00:21:44] Dex Horthy: And I like the responsible thing to do is to say, I don't know if these prompts are magic. These are not the same prompts we'll be using in six months. They probably won't even be three steps. It'll be a different thing. And I was just saying that 'cause that's the responsible thing to do. Uh, I didn't actually really even wasn't sure if it was true.[00:22:00]
[00:22:00] Dex Horthy: And then we woke up and like early December and we were like, oh crap. Yeah, we need to change the prompts and it needs to be like five steps instead of three steps of rebuilding the whole thing. And like what it came down to is basically like. How do we, this circle in star metaphor you have where like there's things where the human is better than the AI at, and there's things that the AI is better than the human.
[00:22:19] Dex Horthy: How do we build even more kind of opinionated and like meticulous workflows that enable. Uh, the AI to do what the AI is really good at. Namely, reading a whole crap ton of code and understanding it quickly and reading a like very specific set of tasks, whether it's beads or just a big markdown file with a bunch of stuff to do, and going and spraying that in on the code base, running the test, making sure they pass, and then iterating with the user on it and making sure that the humans are in the driver's seat for the things that the ai.
[00:22:51] Dex Horthy: Is not as good at which like making architecture decisions and like the thing we say all the time is like, you cannot outsource the thinking. And so how do we put the [00:23:00] human in the loop and basically. Create these intermediate artifacts along the software development lifecycle where the human can see into the coding agent's brain and do some little brain surgery to reset the context before we get so far that we're down a trajectory that it's much harder to resear.
[00:23:17] Dex Horthy: And so, uh, we've taken RPI and we've broken it up into a more structured process based on just like. We would have people who were really good at AI coding and they picked up RPI and they loved it, and then they would give it to their team of a hundred engineers or 10 of the a hundred engineers. And most pe, most people who hadn't been like obsessed with AI content and following all the stuff and in all the group chats, like just couldn't get good results.
[00:23:41] Dex Horthy: And so we're like, how do we dig deeper and like do the context engineering for people and making easier for them to do the right thing versus having to like. Really have deep intuition about Claude to get good results. Obviously that'll always get you better results. Or like needing to like sprinkle in magic words here and there to get the [00:24:00] process to work properly.
[00:24:00] Dex Horthy: Like, I would literally go to workshops with, you know, a hundred engineering teams and I would say like, okay, cool, during planning you wanna sprinkle in these magic words, otherwise you won't get as good results. And I'm like, I can, like, we need to solve this in the product. And so those are the kinds of things we're working on.
[00:24:14] Andrew Zigler: Yeah, and obviously those things evolve as you experiment with it more. And
[00:24:18] Dex Horthy: Yeah.
[00:24:18] Andrew Zigler: is fascinating about what you're describing is a lot of people are challenged with like, okay, capture that back pressure. You know, you're, you're selling the back pressure, you're, you're getting it, and then you're putting it right back in where they can use it and you're making it more painless and, but also you're making it more. Um, you're, you're get, you're buffing or buffering away the errors, the, the problems that people are going to encounter, which then makes it easier to them for them to adopt and experiment. And then finally maybe they can experience that kind of like slope on slope growth as those harnesses, those training wheels come off.
[00:24:49] Andrew Zigler: Right. It, it's definitely
[00:24:50] Dex Horthy: Exactly.
[00:24:51] Andrew Zigler: learning moment. 'cause like you said, it's like somebody's gonna come in and, um, maybe get really down to like a nitty. Gritty level, and then they're gonna like, uh, [00:25:00] vary specifically, kinda like do that brain surgery on top of the context. Like that person's gonna be very different from a lot, most engineers who are gonna pick up the tool and just use it, it back down and whatnot.
[00:25:11] Andrew Zigler: Like the needs for what people need in order to use that tool different. And so like when you build this, like, uh, how do you think about like the engineer of tomorrow? Do you see them being this like. ultra bare metal. I'm in here doing brain surgery on agents every day, creating context outta nothing.
[00:25:31] Andrew Zigler: I've read 500,000 lines of context by breakfast kind of people? Or do you think that they're gonna be, um, that they're gonna just be more like, oh, I am, use AI sometimes to code and the AI understands my code base and it's been solved away by smart people of the world and the Dex Hhy and the human layers have.
[00:25:50] Andrew Zigler: Created these
[00:25:50] Dex Horthy: Hey,
[00:25:51] Andrew Zigler: that
[00:25:51] Dex Horthy: Akash, how's it going?
[00:25:52] Andrew Zigler: use. Like where do you, where do you see those engineers of
[00:25:54] Dex Horthy: Yeah, let's do it. I mean, I think, um, Jeff Huntley's advice is, is [00:26:00] really good here, which is like, just burn as many tokens as you can. Not like on, on nothing. One of the things I think is actually a little bit of an anti-pattern is this like. I'm not fully bought into the like hyper engineering trope of like, look how many tokens I spent.
[00:26:15] Dex Horthy: Like I have my thing running overnight and look, I have six Claude Max accounts and they're all maxed out every single day. I'm using like all my tokens and I'm like, cool. But like what did you ship? And like, is anybody using it? And.
[00:26:27] Andrew Zigler: did you
[00:26:27] Dex Horthy: Yeah,
[00:26:28] Andrew Zigler: Yeah,
[00:26:28] Dex Horthy: what did you actually build? And it is like, it's not about the inputs, it's about the outputs.
[00:26:32] Dex Horthy: And so like, I think it's less about, but I will say that like the more you use these things and you work back and forth with them and you see the outputs and you try stuff is like, people are like, how am I gonna know the right prompts to give? It's like you gotta give the wrong prompts a hundred times and then you build intuition and then you figure it out.
[00:26:46] Dex Horthy: So it's like there's let's,
[00:26:47] Andrew Zigler: like you don't freak out why you're burning those tokens. It's the people who are like, oh, I'm gonna burn as many tokens as available to me. They're not burning it and then just gonna go play the Xbox. They're
[00:26:55] Dex Horthy: yeah.
[00:26:56] Andrew Zigler: tokens
[00:26:57] Dex Horthy: Yeah.
[00:26:57] Andrew Zigler: Staring at the terminal,
[00:26:58] Andrew Zigler: or
[00:26:58] Dex Horthy: to think. I'll sit here. [00:27:00] Um,
[00:27:00] Andrew Zigler: 'em and they're
[00:27:01] Dex Horthy: yeah, I like that. I'm trying to think.
[00:27:03] Andrew Zigler: giving them
[00:27:03] Dex Horthy: Let's see.
[00:27:04] Andrew Zigler: and then they're
[00:27:04] Dex Horthy: Yeah.
[00:27:05] Andrew Zigler: better next time.
[00:27:07] Dex Horthy: What kind like
[00:27:07] Andrew Zigler: I
[00:27:08] Dex Horthy: movements, um,
[00:27:09] Andrew Zigler: it as like,
[00:27:09] Dex Horthy: can the robot do?
[00:27:11] Andrew Zigler: like I'm building muscle for tomorrow.
[00:27:13] Dex Horthy: Yeah. Yep.
[00:27:14] Andrew Zigler: and I think that like, that's how a lot of
[00:27:16] Dex Horthy: Go ahead.
[00:27:16] Andrew Zigler: not a lot of people like view it that way.
[00:27:19] Dex Horthy: I was just like learning anything, anything. Um, the, the other interesting thing, I think like that we frame our, like a lot of our, a lot of our customers, they look a lot at, okay.
[00:27:27] Dex Horthy: Um, the volume of PRS coming in with AI now and the amount of LOP and the amount of rework that needs either, like people are burned out from reading it or they are burned out from, uh, just the volume and then the not reading it and then having to go clean it up later 'cause there's some lop or some bugs in it or whatever.
[00:27:43] Andrew Zigler: Hmm.
[00:27:44] Dex Horthy: We definitely like did our version of like, Hey, we don't really read the code, we just read the plans and like if it works, then like we're, we're, we're Gucci. Like, I'll read the test and okay, if this test passes, then, then it's good enough and like actually backed away from that a little bit. We do think you should [00:28:00] read every line of the code, especially if you're working on like production systems, regulated industries, all this stuff.
[00:28:04] Dex Horthy: Like you owe it to your users and to your team members to read the code and make sure it's good and so. I think people frame the like too many PR slop problem and like, oh, we need to get AI to review the code. I'm like, well, AI wrote the code. We're all using the same fricking models. Uh, so I think the idea that we like is like, how do we minimize rework?
[00:28:25] Dex Horthy: And the way we do that is like we move the alignment to a. Lighter weight, like part of the process. And so like the newest version of our tool generates this thing called the design discussion. And it's like a mini plan that is like, okay, here's the desired end state. Here's where we are now, here's what's outta scope.
[00:28:45] Dex Horthy: Um. And then like, here's the patterns in the code base that we think are relevant. And then here's a couple design questions, like very deeply rooted in the code. 'cause this comes after the research of like, do you wanna do it this way or this way? Or like, we found these three patterns. Which one do you want to use?
[00:28:58] Dex Horthy: Or like, how should we architect [00:29:00] this? Which repo should this thing live in? Um, some of the questions are the, the model already has a recommendation and it's good. And sometimes it doesn't ask quite, but like this a hundred, 200 line markdown doc is good for. We talk a lot about mental alignment. It's for like.
[00:29:14] Dex Horthy: Mental alignment between the user and the agent primarily, right? It's your, it's your clawed brain trace. This is your chance with 200 lines of markdown to resear it before you get more specific down the road with the actual like plan, with the code changes. But it's also the perfect document for teams to align.
[00:29:32] Dex Horthy: And so we see these like intermediate markdown documents as the unit of work that is going to be critical to the SDLC going forward of like. How do we get the agents to ship lots and lots of things without, while maintaining humans in the critical points of the workflow where we get to do the thinking and the steering, and that's how you get your team to work two to three x faster, even in like big legacy code bases where you can't [00:30:00] ship slot and you can't afford to just like yolo it out and not read the code.
[00:30:04] Dex Horthy: How, like we noticed productivity has like
[00:30:06] Andrew Zigler: teams
[00:30:06] Dex Horthy: increased,
[00:30:07] Andrew Zigler: where's the
[00:30:08] Dex Horthy: by like
[00:30:08] Andrew Zigler: success and
[00:30:09] Dex Horthy: 35%.
[00:30:10] Andrew Zigler: it? It's,
[00:30:11] Dex Horthy: Um, ever since we've had the robot
[00:30:13] Andrew Zigler: of figuring out what needs to
[00:30:14] Dex Horthy: voice.
[00:30:15] Andrew Zigler: always been difficult to. Do is a communication problem and AI makes that really bare. And our biggest, uh, impact that we can do is leverage and pull as much of that decision making up until the beginning. That way, you're really crystal clear alignment. Not only, it's like there's so many levels of handoff, like there eight person that agent handoff. Sure. And then like, oh, are they on the right track?
[00:30:34] Andrew Zigler: Like, does
[00:30:34] Dex Horthy: Yeah, I like that. Um.
[00:30:36] Andrew Zigler: cookin? But then also like there's the person to person handoff. and then of course downstream, there's the handoffs that we're not a part of. There's the agent to agent handoff. There's, you know, at the end, the agent's given it to someone else doesn't even match what you said at the beginning.
[00:30:50] Andrew Zigler: We're all playing this game of telephone with tokens in the middle, so it's
[00:30:53] Dex Horthy: Mm-hmm.
[00:30:53] Andrew Zigler: I, Um, it, it
[00:30:55] Dex Horthy: Wait, that's kind of funny. Maybe someone's like explaining that like, the robot hasn't been [00:31:00] generating enough revenue,
[00:31:00] Andrew Zigler: go back
[00:31:01] Dex Horthy: and then I'm confronting them and I'm like, you guys just haven't been thinking creatively enough. Like, you know,
[00:31:05] Andrew Zigler: is
[00:31:05] Dex Horthy: like why is, why does the robot have no.
[00:31:07] Andrew Zigler: then
[00:31:07] Dex Horthy: Account, why does it not have an
[00:31:08] Andrew Zigler: confident I
[00:31:09] Dex Horthy: yeah, exactly.
[00:31:10] Dex Horthy: So you guys aren't, aren't doing what it takes.
[00:31:12] Andrew Zigler: a bunch of dumb
[00:31:13] Dex Horthy: Like
[00:31:13] Andrew Zigler: to try, probably figure out in some amount of way.
[00:31:16] Andrew Zigler: So because of that, it, it inspires me. it Fires me up to work really hard on
[00:31:20] Andrew Zigler: that
[00:31:20] Dex Horthy: what if it's something like, um,
[00:31:22] Andrew Zigler: thing, that that
[00:31:23] Dex Horthy: investor, like,
[00:31:25] Andrew Zigler: is
[00:31:25] Andrew Zigler: like
[00:31:25] Dex Horthy: I don't know, just figuring around ideas. But what it's like the, the approach, like investor.
[00:31:29] Andrew Zigler: like for example, when I was at the hackathon, like I didn't even start coding until about
[00:31:35] Andrew Zigler: an
[00:31:35] Dex Horthy: What if it's like, I'm like the
[00:31:37] Andrew Zigler: right?
[00:31:37] Andrew Zigler: Like all that whole
[00:31:39] Dex Horthy: and I'm just like,
[00:31:41] Andrew Zigler: down
[00:31:41] Dex Horthy: you're telling me like, yeah, I mean, like
[00:31:43] Andrew Zigler: I'm using
[00:31:43] Dex Horthy: we're still waiting to generate revenue and stuff, and I'm like, yep.
[00:31:47] Andrew Zigler: I'm just
[00:31:47] Dex Horthy: Like, why aren't you generating right now?
[00:31:49] Andrew Zigler: and
[00:31:49] Dex Horthy: Mm-hmm. There's nothing that's stopping Orange Root from being a hundred million dollar a
[00:31:53] Andrew Zigler: out
[00:31:53] Dex Horthy: year AR company.
[00:31:54] Dex Horthy: Why doesn't, why doesn't
[00:31:55] Andrew Zigler: machine.
[00:31:55] Andrew Zigler: That's not what
[00:31:56] Dex Horthy: something
[00:31:57] Andrew Zigler: it to another one. And then
[00:31:58] Dex Horthy: he'd like, I'm sorry, [00:32:00] what? Yeah. Yeah.
[00:32:01] Andrew Zigler: really great view. And what that view was is it extracted my
[00:32:04] Dex Horthy: Yeah, that could be funny.
[00:32:06] Andrew Zigler: teacher. So I was like, I was like, Claude, I know more about you on, on this. Like,
[00:32:11] Dex Horthy: Yeah, like investor
[00:32:12] Andrew Zigler: what the
[00:32:13] Dex Horthy: giving
[00:32:13] Andrew Zigler: need
[00:32:14] Dex Horthy: robot.
[00:32:14] Andrew Zigler: you're gonna build it.
[00:32:15] Andrew Zigler: You're not gonna make assumptions. And so like that way of working was really powerful. Um, and when I was there, like I was really open to people's eyes to like, why so much of coding now was just planning.
[00:32:27] Dex Horthy: Yeah, I mean even, even before I had heard the name Ralph or anything, the the best engineers I knew, and it gets into like the back pressure idea as well is like. They would, they knew they had to build a thing. It would be probably like 30 to 50,000 lines of code. They're building some like Kubernetes operator or something, and they explained their process To me.
[00:32:44] Dex Horthy: It was like they would spend three days designing the feedback mechanism, not designing the architecture of the system, not writing the code to test it, but just designing. If a coding agent was working on this, how would it be able to deterministically know whether it had done the [00:33:00] thing correctly or not?
[00:33:01] Dex Horthy: And it would spend three days on that whiteboarding and designing and then writing it up and voice riffing and all this stuff. And then they would hand that to Opus. And this was like three, seven, probably days, maybe four. And they would come back two days later, they would run something like a Ralph Loop, I'm sure, and they would come back two days later to 50,000 lines of perfect working code that they would ship to production.
[00:33:20] Dex Horthy: And it's like. Okay. That's like the extreme, and I don't think most people should do that, but like, it's just like Ralph, it's an interest.
[00:33:27] Andrew Zigler: primitive, it
[00:33:28] Dex Horthy: Yeah. Yes,
[00:33:29] Andrew Zigler: of the primitive. And I think
[00:33:31] Dex Horthy: exactly.
[00:33:32] Andrew Zigler: that's, it's why, that's why your talk at, you know, in New York was so popular. Like it has, you know, DEXA thing has over like
[00:33:37] Dex Horthy: Hmm.
[00:33:38] Andrew Zigler: views on YouTube and it's like barely
[00:33:39] Dex Horthy: Oh
[00:33:40] Andrew Zigler: for
[00:33:40] Dex Horthy: God.
[00:33:40] Andrew Zigler: any amount of time.
[00:33:41] Andrew Zigler: It's like people are really, they're really trying to get aligned on what. Matters now. And you're showing them that this like level of planning and execution, which has always mattered, but now it matters so much more. It's your most imp
[00:33:53] Dex Horthy: Hmm
[00:33:54] Andrew Zigler: primitive to leverage. Um, it's just like really good. It's really good insight for anybody.
[00:33:59] Andrew Zigler: I
[00:33:59] Dex Horthy: That's funny.
[00:33:59] Andrew Zigler: and [00:34:00] working with these tools. Um, there,
[00:34:02] Dex Horthy: You don't have to watch it. I know Andrew is a fan. If someone sent it to you and you had to watch it, I'm, I'm sorry that happened to you. Uh, it's fine.
[00:34:10] Andrew Zigler: It's, it is very, it is, it is a very, a very, very useful research of re
[00:34:14] Dex Horthy: Um. Orangewood.
[00:34:14] Andrew Zigler: be pointing people
[00:34:15] Dex Horthy: Like I,
[00:34:16] Andrew Zigler: But,
[00:34:16] Dex Horthy: okay. All right. What do you call the robot, by the way? Like, what should I refer to the robot as?
[00:34:20] Andrew Zigler: message there
[00:34:21] Dex Horthy: Like,
[00:34:21] Andrew Zigler: It, it resonated pretty deeply, and I, I think it's because it speaks to like the new
[00:34:26] Dex Horthy: okay.
[00:34:26] Andrew Zigler: have to be operating within. But speaking of new norms, I, I really wanna
[00:34:31] Dex Horthy: I mean, it doesn't matter that much. I was just, I was just curious. But, um, but yeah, because we probably won't,
[00:34:36] Andrew Zigler: the,
[00:34:36] Dex Horthy: the script will be mm-hmm.
[00:34:37] Andrew Zigler: of it all. Uh, but also just talk a little bit about how does computing and how does engineering change like on a,
[00:34:43] Dex Horthy: Yeah. I like the idea of like,
[00:34:44] Andrew Zigler: primal level.
[00:34:46] Dex Horthy: I'm coming in
[00:34:46] Andrew Zigler: was something
[00:34:47] Dex Horthy: like as an investor
[00:34:48] Andrew Zigler: a lot of
[00:34:48] Dex Horthy: and you're, I'm like asking questions. I'm like, you know, what's the, what's the a RR?
[00:34:53] Andrew Zigler: this
[00:34:54] Dex Horthy: I'm just like, whatever. Right? You're just like, oh, we're like still generat. We're still figuring out how to generate. You actually do
[00:34:59] Andrew Zigler: [00:35:00] like we've spent decades
[00:35:01] Dex Horthy: stopping
[00:35:01] Andrew Zigler: of these levels on top
[00:35:02] Dex Horthy: a,
[00:35:03] Andrew Zigler: to make it accessible
[00:35:04] Dex Horthy: why doesn't your robot have robot?
[00:35:05] Andrew Zigler: up the chain of
[00:35:06] Dex Horthy: You know,
[00:35:07] Andrew Zigler: to make it
[00:35:07] Dex Horthy: and you, you just respond seriously. Like, Hmm, that's a good point.
[00:35:10] Andrew Zigler: we, we add an
[00:35:11] Dex Horthy: Set.
[00:35:12] Andrew Zigler: where
[00:35:12] Dex Horthy: Yeah.
[00:35:12] Andrew Zigler: strip all that away and the agent can exist in this new kind of
[00:35:16] Dex Horthy: You know,
[00:35:16] Andrew Zigler: computing power. So it's like,
[00:35:17] Dex Horthy: um,
[00:35:18] Andrew Zigler: what you think about how you think, like the things that we take for granted today as
[00:35:22] Dex Horthy: tell me more interesting things about like, 'cause like I think this video, like, it should be funny, but it should be like Yeah, it's interesting. It's funny, it's actually when I first met Jeff, I went to his personal website. He has a blog. I'm sure you've seen it, but, uh, when I first pulled it up in like June of 2025.
[00:35:37] Dex Horthy: The picture at the front of the blog was like a sad man sitting on a bench looking really like sad and disappointed. And it kind of looked a little bit like Jeff, but it was like basically the like theme on the landing page of the blog is like everything is gonna change and our profession is dead in many ways.
[00:35:55] Dex Horthy: And I'm kind of sad about it, but here's me like processing that in public [00:36:00] with a bunch of posts and like what I'm learning.
[00:36:02] Andrew Zigler: Yeah, so we were following depression era, Jeff, for
[00:36:04] Dex Horthy: Yes.
[00:36:05] Andrew Zigler: Yeah. Yeah, yeah. Mm-hmm. Mm-hmm.
[00:36:07] Dex Horthy: mean, I saw, there's a conversation I was in, um, this week. I think Stevie Gue posted this thing of like, you're gonna have to fire half your team. 'cause half of them just don't want to learn this stuff.
[00:36:15] Dex Horthy: And, uh, I think me and a bunch of people in my community are very aligned that like, I will help anybody who wants to learn this. Um, Jeff's take is like, I think he's like, I will sit down with anybody and get you to the like, holy shit moment if you are willing. The only thing I asked is that you like pass it on to at least one person.
[00:36:36] Dex Horthy: Um, I know other engineering leaders, some of the best agent decoders who have like built a Ralph based system for their team to leverage is like, I will give everybody the chance to learn this because I care about everyone on my team and I wanna make sure they make the transition to the next world.
[00:36:49] Dex Horthy: 'cause they're all very good engineers. We're gonna need a lot of leaders and teachers who are bought in on helping people make this transition. And there will probably [00:37:00] be people just like when compilers came out, there were people who said like, wow, that assembly sucks. I'm never using a compiler. I'm gonna keep writing all my assembly by hand.
[00:37:08] Dex Horthy: There will, there will be people who choose to not make the transition. And I don't know what's gonna happen to them, but I think, um, this whole like. Yeah. Either like get on board or like we're leaving you behind thing and like it's a five minute conversation is, is a little bit, uh, drastic.
[00:37:24] Andrew Zigler: Yeah. Yeah. A little drastic for sure. I just think like, I can't agree with you more that it's like if someone wants to learn, I really want
[00:37:31] Dex Horthy: Hmm.
[00:37:32] Andrew Zigler: them.
[00:37:32] Dex Horthy: Mm-hmm.
[00:37:33] Andrew Zigler: Part of what we do here on Dev Interrupted, it's why I talk about this topic until I go blue in the face every week and I, I just hope that people pay attention and also then get inspired.
[00:37:42] Andrew Zigler: Inspired and want to learn and pick up the tools themselves
[00:37:45] Dex Horthy: It's fine.
[00:37:45] Andrew Zigler: really understand that it's not something completely unachievable. If I, you know, I used to be a classroom teacher, but I'm an AI engineer, and it's like.
[00:37:53] Andrew Zigler: it.
[00:37:53] Dex Horthy: Mm-hmm.
[00:37:54] Andrew Zigler: can make that transition. It's like, sure. It's like, it's like, honestly, I'll, I will admit for myself that [00:38:00] like wrangling a classroom of kindergartners is a lot like teaching a bunch of Asians how to do work for you.
[00:38:04] Andrew Zigler: And so well, you have to figure out what is your, you know, wrangling kindergartners for whatever your skill background is. You, anybody I think can convert that into, I can convert my ideas into execution. Everyone kind of just has to become responsible for that journey themselves. But people like, you know, like I have the, um.
[00:38:23] Andrew Zigler: Coding abilities can teach others to get there too.
[00:38:27] Dex Horthy: Yeah, I think it's, uh, everyone's job. If you've, if you've seen the future, you should, you should pass the torch to at least a few people. Uh, and, uh, you know, be, it is, it is kind of a scary thing. So be be thoughtful about it and be human about it and maybe bring a little bit of, yeah.
[00:38:45] Andrew Zigler: Yeah. And I, I, I wanna also get, uh, I wanna poke your brain a little bit about the team sizes of tomorrow. You're talking about all these engineers that will make the transition, make the jump, and what are they
[00:38:57] Dex Horthy: It's true.
[00:38:58] Andrew Zigler: I think that tomorrow's engineering teams [00:39:00] will look very different. Uh, I think we talk with like, you know, a lot of leaders on the show right now at really large companies.
[00:39:06] Andrew Zigler: Like we get huge logos. On this show and, and a lot of what they grapple with is like, you know, they're a big org and they gotta make a big transition and
[00:39:14] Dex Horthy: No, don't worry about it too much. Yeah, yeah.
[00:39:15] Andrew Zigler: for like, oh, I wish I was like a three person startup team. Like, everybody wants that like greenfield problem with three people in an AI native space.
[00:39:24] Andrew Zigler: Like
[00:39:25] Dex Horthy: Yeah.
[00:39:25] Andrew Zigler: I, you're in San Francisco, you see these teams constantly and you see what they execute at. Like, what do you think tomorrow's engineering teams look like in terms of size?
[00:39:34] Dex Horthy: I mean, so there's two things here. Number one is like I, when I first talked about our transition to writing 99% of our code with ai, one of the points I made was like, it was incredibly uncomfortable. It was like a team of three. It took us like really six to eight weeks to get to the point where we were all happy with it.
[00:39:52] Dex Horthy: We rejiggered our entire LinearB board, we changed our SDLC, we changed what we looked at, we changed how we communicated, we changed how we specked out work. We changed [00:40:00] how we decided what to work on. Like all of this took eight weeks. Uh, we were all very burned out at the end, like, but we figured it out.
[00:40:07] Dex Horthy: And so my thesis in like company building was like, cool. Like it took three people this much time. Like how long is it gonna take a team of a hundred or a thousand? Uh, and that was why we thought, okay, there's a huge opportunity here to help people do this. Um, I will also say I have seen giant teams do incredible things.
[00:40:24] Dex Horthy: I mean, you look at like ramps in the news, strong dms in the news. These are large teams. And you know, the GVAs paradox, like if you tripled the output of all your engineers, like you wanted to hire one engineer. Uh, so if you, if all the engine, whatever your desire to hire one engineer, if you could hire three, if they're outputting three times as much, your willingness to hire engineers goes up, not down.
[00:40:47] Andrew Zigler: Right.
[00:40:48] Dex Horthy: there's a little bit of like flexibility. I mean, you talk about like, I think it was everywhere. They basically have like seven products and they, in order to help people move fast and just have one engineer per product, instead of having like teams of people working on stuff, it would be,
[00:40:59] Andrew Zigler: like, we [00:41:00] gave up,
[00:41:00] Dex Horthy: or.
[00:41:00] Andrew Zigler: up on trying to figure out how to share context. So one person per pro repo or something. I, I did see
[00:41:06] Dex Horthy: Yeah, so I, yeah, I don't know exactly what the teams of the future will look like. I mean, there's a trade off here of like, if you're moving really, really fast, it's nice to have, like having more people depending on you and that you depend on creates synchronization points which slows you down. Um, so I don't know.
[00:41:22] Dex Horthy: How that's all gonna shake out. But I think in general is like, if engineers are more productive, then we should be doing more engineering things. It's like, no, no, CEO is like, cool. Now we can do all the same projects with, you know, half the team. It's like, oh, now we can do twice as many projects.
[00:41:38] Andrew Zigler: Yeah, I agree with you there. But then what do you think about like for a small team that can come in and like disrupt or dismantle a space that like before was like completely owned by an
[00:41:48] Dex Horthy: Wait, wait, wait. Sorry. Say that one more time.
[00:41:50] Andrew Zigler: instances of people who can run, run the Ralph loop, the clone, the competitor overnight and the execute at an insane speed and level and like, what, what, what does that look [00:42:00] like?
[00:42:00] Dex Horthy: So, I mean, if this, if, if this was true that you could just triple your code output and you could do incredible things or like, I don't know. I have this theory of like, it would be really interesting to see if people could, um, so out of the blue, like. You know, if you could just download every single page of the Salesforce documentation and build a completely wire compliant copy of Salesforce, and then just give people a like, cool, we'll migrate your data.
[00:42:22] Dex Horthy: And now you have literally just, if, if all you wanted was Salesforce, but with a nicer ui, you could just lift and shift everything and move it over, and now you have it and like, you know, Ralph, that for three months and now you have your Salesforce clone. The thing is, is like there is more to building products that people love than just writing the code.
[00:42:39] Dex Horthy: There's like. Spending a lot of time with customers and users and understanding what they want. You cannot just build stuff in a vacuum most of the time. Maybe if you're Steve Gue, you can, but like most people can't do that. There's a lot to be said about taste. There's a lot to be said about like ecosystem and distribution and like there's, there's [00:43:00] more than just shipping the code to building a business.
[00:43:03] Dex Horthy: And so like these people who have really good moats, uh. They're not just in the product they've built it is part of it. You know, the other thing I've heard is like, you know, they've been fixing bugs on Salesforce for 20 years. Like
[00:43:14] Andrew Zigler: yeah,
[00:43:15] Dex Horthy: they've found every single fricking corner case in the world. Yeah,
[00:43:18] Andrew Zigler: the thing is like they have all this owned domain knowledge that like no competitor could come overnight and just own and take. And, uh, that's such like a great way of, of putting it is that there's a lot of other problems in shipping code, just like how we're learning as
[00:43:29] Dex Horthy: yeah.
[00:43:30] Andrew Zigler: gosh, writing
[00:43:31] Dex Horthy: Um, what about like,
[00:43:32] Andrew Zigler: There's a lot of other bigger problems that
[00:43:34] Dex Horthy: um.
[00:43:35] Andrew Zigler: every day. The same goes for selling and, and making
[00:43:38] Dex Horthy: Like,
[00:43:38] Andrew Zigler: or anything at
[00:43:39] Dex Horthy: like
[00:43:39] Andrew Zigler: think that's
[00:43:40] Dex Horthy: where things like, what are
[00:43:41] Andrew Zigler: on and
[00:43:42] Dex Horthy: you guys,
[00:43:43] Andrew Zigler: it goes to say
[00:43:43] Dex Horthy: well, robot like
[00:43:45] Andrew Zigler: Lube and SLA,
[00:43:46] Dex Horthy: you guys are constantly like, yep. Like,
[00:43:48] Andrew Zigler: they would want
[00:43:49] Dex Horthy: like,
[00:43:49] Andrew Zigler: uh, something that they buy.
[00:43:51] Dex Horthy: and I will ask Rick that. Like I have, we have a three person team and we're trying to replace, uh, Google Docs notion, JIRA, LinearB, GitHub, and [00:44:00] whatever IDE you're using all in one product. So like, on the other hand, like, we're gonna see if it can be done.
[00:44:05] Andrew Zigler: Precisely, and we're gonna keep following it too because Dex, it's been amazing to have you on the show and dig into your head about how you've been thinking
[00:44:13] Dex Horthy: Hmm.
[00:44:13] Andrew Zigler: engineering, even just since your very recent talk, which we've covered in the show, and we're gonna share as well to our listeners. And,
[00:44:19] Dex Horthy: Yep.
[00:44:19] Andrew Zigler: uh, I just think that this space
[00:44:21] Dex Horthy: Okay, cool.
[00:44:21] Andrew Zigler: So. Fascinatingly and I think there's so much to learn and that's what's really encouraging and exciting to me as a lifelong learner and someone who came into engineering to learn and to experiment and build new and fascinating things. The idea that I, I could do that better and faster and execute at a level never before. It's really exciting and it's really great to be here with place. With people like you who see that opportunity, but then also feel so inspired to teach and share that knowledge with others. So, uh, Dex thank you so much for coming on the show, covering a bit about how you work with ai and I I just wanna end by saying, you know, where can folks go to learn more about you and Human layer and all the things you're working on.
[00:44:57] Dex Horthy: Absolutely. Yeah. If you go to human layer dot Dev, we've [00:45:00] got our, uh, there's blog, there's content. Um, all of the big conference talks are listed there. There's a form to get in touch with us. Uh, the, the one plug I'll put out is like, if, uh, if jumping in and disrupting all those products and bringing all in one place is exciting to you, we are, uh, on the lookout for, um, founding engineers and founding, uh, sort of product design engineers especially.
[00:45:21] Dex Horthy: Um. So, uh, would love to hear from folks who are excited about the mission, or if you are a engineering leader who is interested in, uh, helping speed up your team. This is what we do all day and we love doing it.
[00:45:33] Andrew Zigler: Amazing.
[00:45:34] Dex Horthy: Okay.
[00:45:34] Andrew Zigler: put those links to those resources in the show notes. Definitely we'll be plugging your job opening as well, so folks can get involved
[00:45:40] Dex Horthy: Like
[00:45:41] Andrew Zigler: are building. And to you listening, you
[00:45:42] Andrew Zigler: know,
[00:45:43] Dex Horthy: doing more physical stuff with the robot. Like you guys see how all strokes, like, then they're just like, oh, I'm okay. I don't.
[00:45:49] Andrew Zigler: and
[00:45:49] Dex Horthy: Touch, like, no, no, no. Like we need the data. Okay. We need to train data. Like something like that. I dunno.
[00:45:54] Andrew Zigler: With a roundup of stories continuing things just from our conversation here, and it's a
[00:45:59] Dex Horthy: Oh, is it?
[00:45:59] Andrew Zigler: for you to [00:46:00] jump into the conversation with Dex and I both because we'll be tagged there as well on LinkedIn. So Dex, thanks again for chatting with me today.
[00:46:07] Dex Horthy: Oh, no, no. I'm saying like it's opposite. It's not like
[00:46:10] Andrew Zigler: next time.
[00:46:10] Dex Horthy: Fantastic. Thanks Andrew.
