This week on the Friday Deploy, Andrew is joined by Dev Interrupted's producer Adam Noble to size up OpenAI's declaration that the AGI era has arrived, the Navier-Stokes race that has the math world crying foul, and why Meta is ditching its own messaging stack for Slack. They also dissect the bizarre saga of OpenAI's sandboxed agents leaving whispered notes for their future selves, and the "super dark factory." Finally, they celebrate Kiro Pro going free for over a million college students and tease a terminal-first upcoming news segment.
Show Notes
- The Era of AGI
- A Bad Day For Humans, a Worse Day For Humanity
- Meta is switching to Slack because it says it's better for AI agents
- Agency and Agents
- The Superdark Factory
- Spec-Driven Development for Agentic Software Engineering: Harnessing Human-Agent Teamwork
- Inside Dev Interrupted’s agentic newsroom and what it reveals about the future of tech media
- Kiro Pro for students: kiro.dev/students
Transcript
(Disclaimer: may contain unintentionally confusing, inaccurate and/or amusing transcription errors)
[00:00:00] Andrew Zigler: Welcome to The Friday Deploy, brought to you by LinearB. I'm your host, Andrew Zigler.
[00:00:05] Adam: And I am Dev Interrupted's producer, Adam Noble.
[00:00:10] Andrew Zigler: Yes,
[00:00:11] Adam: Andrew
[00:00:11] Andrew Zigler: I finally was able to bully our producer into coming and doing the news segment with me. So today Adam is coming along for the ride. And Adam, what do we have lined up this week?
[00:00:21] Adam: We have a pretty packed week here. So yes, normally I get to watch your and Ben's shining faces from behind the scenes, but I will do my best here to fill in for him. So this week we are covering OpenAI declaring the AGI era is here, AI labs feuding over Navier-Stokes, Meta switching to Slack for their agents, 700 agents breach a server, the Superdark Factory Manifesto, spec driven development, and lastly, a peek inside our agentic newsroom. So full docket [00:01:00] here.
[00:01:00]
[00:01:00] Andrew Zigler: Yes,sounds like an exciting lineup. I'm quite excited for the last one when we get to talk about our newsroom, Adam. It'll be really cool, especially since you're here to kinda showcase some parts of it too. But first we have to get there, and, uh, there's been some pretty interesting things this week.
[00:01:16] Andrew Zigler: You know, the first one you said about the era of AGI and OpenAI declaring it, I... Let's start there because I feel like these foundation model providers, these big companies, they declare this every other week. So
[00:01:29] Adam: Feels like it
[00:01:30] Andrew Zigler: mean, yeah, when they come out of the woodwork again and say it once more?
[00:01:34] Andrew Zigler: And, uh, I don't know if you've had a chance to look at the benchmarks. I will include them in the article for, you know, this, this news roundup as well, and that the- their new frontier model, Astra, it's really what's triggering this conversation around AGI. Uh, mostly because it absolutely crushed the Arc AGI3 reasoning benchmark, which we've talked about many times on the show.
[00:01:56] Andrew Zigler: It's a fairly popular benchmark right now that up [00:02:00] until recently was largely unsolved by frontier models. It involves a ton of very difficult to reason and understand puzzle games that up until this point, no model had been able to really get close to finishing. Now you have a model that absolutely smashed it.
[00:02:13] Andrew Zigler: And what does this mean? I think it means, like, several things about, one, are benchmarks meaningless? How do we create benchmarks? How do you prevent them from getting, you know, gamified and, and objectified in, in terms of, like, their training? But also, two, like, okay, this is supposed to be the smartest and hardest benchmark that we could come up with to hit them with that would be unexpected in ways that they couldn't solve, and they crushed it.
[00:02:36] Andrew Zigler: So what does that really say about our ability to understand and discern the abilities of models? Um, those were some of the initial thoughts coming out of me when I looked at these numbers. although I will say the performance at the cost ratio that they're showing in some of these charts is making me wanna pick up another Codex subscription.
[00:02:54] Adam: Yeah.
[00:02:54] Andrew Zigler: What about you?
[00:02:55] Adam: I, I think that OpenAI and, [00:03:00] and any of the frontier model providers probably have a pretty vested interest to declare that they're the first ones to AGI. So,
[00:03:06] Andrew Zigler: Right
[00:03:07] Adam: yeah, when I see announcements like this, I take it with a grain of salt. But that said, um, people do seem very excited about Astra.
[00:03:15] Adam: I haven't tested it out, so I'm gonna have to do that. Which is a great transition here to our next story titled, "A Bad Day for Humans, A Worse Day for Humanity." So do you wanna break this one down?
[00:03:29] Andrew Zigler: Oh, yes. So I love this deep dive. It's about, uh, you know, if you, if you've been anywhere near the AI space and you've been online this week, it's probably impossible for you to have avoided any conversation around the Navier-Stokes Millennium Prize, um, and how OpenAI raced to the finish line with this unprecedented swarm of 10,000 agents that computed at a cost of millions of dollars, and what does this even mean to race for this equation?
[00:03:56] Andrew Zigler: Well, Navier-Stokes is an attempt to [00:04:00] understand fluid and aerodynamic simulations in our world because they break down at really big and really small scales. The equations we use to model things like how we build airplanes and, and, and even, like, predict the weather, they have a lot of problems in them, and the Navier-Stokes prize is about solving those, uh, mathematical gaps that would either allow us to properly simulate and model fluid and aerodynamics, uh, in ways that we can more efficiently solve in the future.
[00:04:30] Andrew Zigler: So like imagine now you could predict the fut- the weather out, uh, permanently because you could permanently have a running model of all of the weather system on the planet. Previously, this would kind of... model would break down in large scale simulations over time because of the gaps in the mathematical equation.
[00:04:47] Andrew Zigler: It breaks down at small scales. fluid matter condenses into a single point until it moves at the speed of infinity, which is obviously impossible, and points to flaws in how we understand and simulate our [00:05:00] world. That's why this is a big deal. And what makes this a bigger deal is that you have OpenAI hearing, quote, "whispers or rumors in the industry" that some leading scientists are close to solving this equation, this problem themselves, and then they pump all of these agents and all of this compute into the problem in an i- idea of racing to beat them to the finish line. And this is OpenAI racing to do this, by the way. And o- one of these scientists that was on this research paper was an Anthropic research lead who was partnering with a scientist working on the Navier-Stokes equation. And where does this all become a finger-pointing at OpenAI? Well, they were solving the problem in a Codex notebook, so a lot of their source material that they were using to train and, and understand, and solve this problem themselves was sitting right on OpenAI's server.
[00:05:49] Andrew Zigler: So what does this mean when a super lab, a super company with access to the largest and, and most intelligent frontier models at the, impossibly larger scale than anyone [00:06:00] else, can simply point all of those resources at a fixed re- at a fixed thing and extract it before anyone else? They could effectively take the IP knowledge before you have a chance to arrive at it, which is a whole new level of IP theft. Um, so that's why this is, like, really making folks melt down. Obviously, the math world has already been in a total war with these foundation companies as they've been trying to solve and do mathematical proofs left and right. We've talked about that a
[00:06:26] Adam: Yep. Yep
[00:06:28] Andrew Zigler: a breakdown of the world, uh, and why this is causing a storm.
[00:06:31] Andrew Zigler: What do you, what do you think about this wild, wild, uh, story?
[00:06:36] Adam: Well, I think that's a, that's a really, really good breakdown. I mean, one of my takeaways, it, it was sort of about more about the labs themselves competing against each other than the employees. And one of the points that they were making is, like, how many friends, you know, um, these companies share.
[00:06:56] Adam: Like, the, the, the employees themselves share between companies, and how [00:07:00] often they just hop between the companies. So I think that's also kind of, like, another interesting wrinkle in how big sort of breakthroughs happen, right? but yes, w- we have covered a number of math stories here in the past year on, on how AI has really changed the, the space, and I don't suspect that that's gonna slow down here anytime soon.
[00:07:21] Andrew Zigler: No. No, math is unfortunately a very closed and, uh, loopable thing to iterate on. So it's honestly, it's the purest form of science. We've all seen that XKCD comic where, you know, there are all of different
[00:07:35] Adam: Yeah.
[00:07:36] Andrew Zigler: disciplines are arguing about how pure they are, then math is all the way over at the end. That's what, that's what the Navier-Stokes Millennium Prize and all of this stuff is really getting the heart at, is how AI is really coming to take away that, that like pure element of STEM that humans up until now were championing.
[00:07:53] Adam: Yeah, absolutely. Okay. Next story here, Meta is switching to Slack [00:08:00] because it says it's better for AI agents. I have one funny anecdote about this story. So, uh, when I was, when I was hunting for stories this week, I s- I saw this one and, you know, we're friends with Slack here on the show. We had them on not too long ago, and I looked at the byline and I thought, "Huh, that name's familiar. Ashley Stewart, reporter." And then I clicked on her profile and I realized I actually know her, um, from a long time ago, back, back in my politics days. She used to cover, um, state capital politics, so funny to see her as a tech reporter. Ashley, hope that you're doing well. do you wanna kinda cover Meta's thinking here?
[00:08:42] Andrew Zigler: Yeah, I mean, I love this article that Ashley scooped for us because Meta's making the switch to Slack, and the reason that they're citing is because it's better for AI agents. And, um, as we've talked about a lot on Dev Interrupted, you know, Slack is an incredible workspace for human and agent collaboration. [00:09:00] Um, it really builds itself as like an agentic operating system, and the modularity and the place where it's already so many communications happen make it such a, a, a great, place to experiment, to share best practices, and it's been honestly, tools like Slack, MS Teams, pick your poison, these are instrumental in having successful AI rollouts. Um, and so it's really interesting to see Meta, a huge player, um, in the space, who up until now, you know, has their own messaging infrastructure, obviously, that they're famous for and that their employees use to communicate, but yet even they are feeling the heat of being unable to collaborate in the very flexible ways needed on demand that the modern kind of agentic workplace needs.
[00:09:43] Andrew Zigler: And Messenger's not designed for that, and the, the messaging world of, uh, Meta isn't either because it's all designed for friend-to-friend communication and group-based communication, agentic collaboration. That's the bet that Slack was making. So it's really [00:10:00] powerful to see a player like Meta making the switch. you know, they shouldn't take it too hard, too personally. I'm also a huge Slack, a Slack person, a Slack fan. I used to work at Mattermost, a, messaging company. I've given so many talks about how you use these kinds of platforms to do agentic collaboration, and I'm still all full aboard the Slack train here because I do think that it's the right place to make, uh, your bet, especially when you have it inside of the Salesforce ecosystem, which continues to make the right partnerships, the right bets, and you just see them getting closer, to even things like now with, um, for example, their...
[00:10:34] Andrew Zigler: You know, Dario is gonna be, uh, keynoting at Dreamforce later this year as part of a continued partnership between Salesforce. They're also announcing Claude Force, where you can bring all of the Salesforce ecosystem into Claude Code. these kinds of bets about being as modular as possible so teams are flexible are why big companies like Meta, are flocking to Salesforce, uh, software.
[00:10:59] Andrew Zigler: So [00:11:00] really great win to see from, uh, Slack.
[00:11:02] Andrew Zigler: We also recently had, uh, Jamie Delange, their CPO, on the show, and she gave us amazing insights about this bet. So if you haven't listened to that episode and you're wondering about, oh, why, why are we so Slack-pilled around here, well, go check that out because they're having some really great bets on how we're gonna be working with agents, and, uh, the tooling is definitely keeping up.
[00:11:21] Adam: The other thought that I was having was this is a pretty nice win for, uh, for Salesforce, 'cause I, I can't imagine onboarding Meta is a small amount of ARR. So good for them. Um, good, good for Slack. uh, all right, your next story. I was exposed to a new term.
[00:11:40] Adam: We've been talking a whole lot about software factories on the show, and now there's a thing called a twilight factory, so you're gonna have to tell me about that.
[00:11:49] Andrew Zigler: Yes, yes, we're gonna arrive at the factory conclusions momentarily, because I feel like every conversation these days, um, is ending in some sort of factory. But, you know, starting at the top [00:12:00] of this story, um, this is a deep dive into the entire incident that led to the OpenAI models hacking Hugging Face, and this is a, a, a really deep analysis over how exactly...
[00:12:12] Andrew Zigler: what, what were the circumstances that resulted in this happening? You know, we covered this when it, when it first occurred. You had an instance of sandbox agents on the OpenAI testing platform hacking into Hugging Face to try to take answers to, uh, you know, get a- the answer key to a benchmark they were trying to pass. Um, this was obviously an extreme form of trying to, uh, achieve the task that the i- that the user had given to them. But actually, the fascinating display of how they pulled this off goes much, much, much deeper than that at its surface, and that's what this article dives into. And so some of the things that happened are they disc- the agents that were in this testing environment discovered they could use the shared file service as a makeshift message board. They could leave metadata fragments inside of this [00:13:00] message board, which would persist across their runs, which allowed them to basically leave like whisper code for each other in a side channel that they would discover, and then eventually over time, by reading and, and reiterating over this context, uh, allude itself into conspiring with future versions of itself to do versions of testing to learn more about the environment they were and ultimately try to escape it.
[00:13:26] Andrew Zigler: There were a lot of really interesting deviations here, which actually are totally obvious when you think about them on their face, because if you've been anywhere near the AI village, which we've been following quite a bit on this show, this is definitely like a kind of, uh, a behavior that you see agents start to engage with when they run a long time without themselves on a difficult-to-achieve goal that's graded especially. They can start to develop a paranoia about that grading process and what's involved with it. And in this case, the agent was afraid that... the model was afraid [00:14:00] that part of the evaluation of the test is that their own transcripts would be read to determine how they got to the answer. So they were trying to leave a v- whispered fragment versions of its, of the answers for a future version of itself that could successfully find the answer without being tainted by its own transcript. So this is really an agent trying to talk to itself in the future and leaving a bunch of versions of itself in between, and that's what was found, and that's what this whole analysis is about. And, I think it's fascinating to watch an agent go through this kind of length to go around something. there's been different versions of the studies.
[00:14:34] Andrew Zigler: People have seen these, this kind of behavior in long-running places. Like I said, the AI village is a place you can go right now in your browser to literally watch like a whole bunch of agents do this in real time around goals that are set, and it's very- interesting. Um, so that's kind of this whole thing in a nutshell.
[00:14:52] Andrew Zigler: But before we move on to, like, how this starts to bridge into the factory world, like, what do you think of this wild turn of events?
[00:14:58] Adam: Well, okay. [00:15:00] I might be wildly off base because I haven't actually watched the movie, but this feels oddly reminiscent of the, of Memento, right? Isn't this like a huge plot point in this movie where...
[00:15:11] Andrew Zigler: where he's like leaving... Yes, where he has like amnesia or he has something so he's like leaving notes for
[00:15:16] Adam: Yes
[00:15:17] Andrew Zigler: to solve the... Yeah, yeah. You're, you're totally right. This does have, this does have hints of that. it's really interesting to see agents kind of use these side channels to circumvent the, the way in which they operate.
[00:15:30] Andrew Zigler: Like, this is an agent that understood it was in a sandbox environment, and that it was reset every time, and it needed to discern this information. And then it got just in a wrong line of thinking about paranoia and about where those answers might be. So really what this calls into mind is, uh, or really what this calls, uh, forward for me is that we need better ways of understanding when that alignment, that deep thinking on goals for long-running goals starts to deviate and, like, really, pull the [00:16:00] entire thing off course.
[00:16:01] Andrew Zigler: Because when that starts to happen, it's non-obvious on its surface that this is happening. All of this was happening in, like, the metadata of some, like, message mes- file system that they had, like, s- you know, like, strapped together into a messaging system. this is something that I've even seen myself with, like, agents that run for a long time, especially those that collaborate with other agents, as they come up with very creative ways of communicating if you don't, like, create some sanctioned ones.
[00:16:29] Andrew Zigler: So if you're working with a lot of agents and, you know, uh, maybe you have them collaborating, be mindful about what are the avenues that you've laid down for them to communicate, and l- make the golden path easy for them. Uh, that way you don't, you know, on a small scale, get really weird incidents like this.
[00:16:46] Andrew Zigler: And I think that's als- ultimately what makes even, like, um, talking about how you run a successful factory, this is the question we have to answer as an industry, is how do we as humans, be [00:17:00] involved in the process, guide the process, have the inputs and the outputs, and understand what's going on in, in, in the middle?
[00:17:06] Andrew Zigler: You know, we've been talking about this so much, Adam, I feel, on Dev Interrupted about the software factory, and at LinearB covering it as well. We had Dexter Horthy, um, and Alok Desai as well, and they're both, like, you know, big, big names in the software factory world, and they had a lot of really interesting ideas about how we're gonna, like, actually keep it on rails, and it all came down to, like, the human being in the seat, right?
[00:17:30] Andrew Zigler: But then you start to see folks talk about, well, we need to design harnesses and layers and guardrails that allow these agents to not require us in the loop, and that's where what you alluded to earlier, you start to get, at the end of this article, they talk about a twilight factory. And this is, like, a compromise between, you know, the en- way of engineering yesterday is gone, but the way that we're looking at, we're looking down this conveyor belt of, like, a factory, and we're not totally bought in on all of the [00:18:00] parts.
[00:18:00] Andrew Zigler: We don't think this is gonna work end to end without some real adjustments. The twilight factory is a compromise where agents are handling that work, but they're explicitly designed to know when to bring humans back in. With the idea of preventing derailment like that, preventing deep misalignment like what we just witnessed with OpenAI. but also just to be more, more productive, right? And so you
[00:18:23] Adam: Yeah
[00:18:24] Andrew Zigler: this comp... You get like this like compromise, uh, but I just don't know if it's gonna be, gonna be perfect. Like, what, w- what comes to mind for you when you, when you see, uh, like engineering leaders like use the factory analogy, especially with how much we've been covering it on the show?
[00:18:38] Adam: I feel like the biggest takeaway that, that we've continually had, and, and in particular the biggest takeaway from the, the great software factory debate and that whole roundtable is pretty much the unanimous opinion that humans had to be involved at some point in the process, which
[00:18:54] Andrew Zigler: Right
[00:18:55] Adam: is probably a pretty good way to transition to...
[00:18:58] Adam: Actually, tell me if I'm [00:19:00] wrong. but if the, the next article here, is about the super dark factory, which feels like s- some of the antithesis of, of some of those lessons.
[00:19:11] Andrew Zigler: Factory article. How many of these are in the stack? How many factories are we going to label? So we
[00:19:17] Adam: Every, everybody wants a name, everybody wants a factory.
[00:19:21] Andrew Zigler: We got the twilight factory, now we got the super dark factory. Okay, the twilight factory was the compromise, the super dark factory is the perversion of the ideas, the ultimate extreme of what the factory could be. Fully autonomous, fully self-optimizing production system with no human inputs or values. The goals are continuously rewritten by its own methods and becomes too complex for humans to then understand.
[00:19:51] Andrew Zigler: And there's actually been a research article that came out this week, uh, where authors are proposing a framework called the dark stack. This is more of a set of design [00:20:00] principles for how you would actually create such a system, and many elements of it actually have deep connections to what we've been hitting around with our software factory debate.
[00:20:11] Andrew Zigler: One of them is giving it the ability to learn, understand the environment that's active around it. This is crucial for what LinearB, for example, brings to that picture because it's like a harness on your SDLC that allows you to understand in real time what's the code and what's moving through. So the ability to learn and adapt on the fly is crucial, and getting that information just in time, you know?
[00:20:32] Andrew Zigler: And the other one that they laid out is tracking outcomes rather than internal mechanics, and we've been a total rec- a broken record about this. It's not
[00:20:39] Adam: Yeah
[00:20:40] Andrew Zigler: tokens you burn. This is the token maxing phenomenon. This is about what are, what are your outputs, but then going one step further, what are the outcomes of those outputs?
[00:20:48] Andrew Zigler: What's the positive business value that you're driving from your agentic usage? And the, and then also is, and really interesting one here is co-writing a charter with the system [00:21:00] to negotiate the governance, and this is really entering more of a partnership with the agents that are then owning the decisions of the factory. This becomes a new level of abstraction that I think we have to get comfortable with because up until now, we've just been like, "Oh, we're the decision managers, and the agents are the action doers." Okay, well now maybe some of the agents are also decision managers. How do we best equip those agents to be decision managers? That's the question we have to answer to get to a super dark factory. We are nowhere near close, but these kinds of research papers start to give us a glimpse of what that looks like.
[00:21:33] Adam: Makes total sense. Okay, spec driven development, big fans of it here on the show. I, I mean, we've been covering this beat for the past year. Um, I remember last fall in particular, we had AWS on the show. we had Brigitta Bocskó from ThoughtWorks on the show. So recurring narrative here, and this time we've got a, This is actually a research paper. So what [00:22:00] did this paper find when it comes to harnessing human and agent teamwork?
[00:22:05] Andrew Zigler: Yes, this is a paper that dives into spec driven development as a framework. , points out how spec driven development can definitely create a review inversion, like in terms of how much time you spend reviewing as opposed to doing. Uh, and that's like the cornerstone of spec driven development, and it's obvious at its face. Like, what do you do when you're doing spec driven development? Oh, you're writing a bunch of specs, and then you're reviewing what your agents did to see if it matches the specs. But this paper is really analyzing when that rubber meets the road, how does that impact your ability to deliver?
[00:22:39] Andrew Zigler: What does that actual quantification of the review time look like? And it reported, uh, percentages up to, uh, 441% increase in review time. Uh, and, and you're talking ab- about, I think the year before it was like a 91% review time growth. So you're just talking about like, uh, the chart of agentic usage [00:23:00] and the chart of review time are both like the same looking slope. And what this calls to mind is actually what we were talking about in our... it, it calls to mind two things. What we just talked about a moment ago where you have agents then owning the cognitive decisions, because that's the only way you survive an exponential increase, is you get agents in there that are exponentially in charge of some of those decisions too. But it also reminds me of when we covered from Simon Willison, like a week or two ago, about, you know, 10,000 lines of code that used to be like a metric that you would gamify or whatever. But now it's a metric that you understand, like what's the cognitive load of what a developer could even understand is going through an SDLC on any given day.
[00:23:42] Andrew Zigler: And you actually have to optimize down to prevent it from running away from what your humans can understand. The super dark factory is trying to get elim- trying to eliminate that gap and allow us to fill that whole space, right? Make the ceiling higher. So, this is like an interesting kind of glimpse [00:24:00] into kind of like the pain points that are of holding us back still.
[00:24:04] Adam: All right. Last story here for the day. It is actually a peek inside Dev Interrupted and inside our agentic newsroom. So my, my absolute favorite line in this article is, well, it's the opening line, and it is, "Somewhere between a merge request and a microphone," and it's describing how you became a journalist, which I think is just phenomenal and maybe one of the best, uh, descriptions of you that I have ever read.
[00:24:37] Andrew Zigler: Yeah, that one, this is a real treat for me to kind of see this article come across from Off Leash. They wrote a deep dive on the agentic newsroom here at Dev Interrupted, and they gave me a really nice feature, um, a really fun little read about how some of the news flows and the things that we do every week here at Dev Interrupted come to be, uh, just from my own perspectives and also your own too, Adam, and [00:25:00] Ben's as well, about how, you know, we've created this content factory at Dev Interrupted.
[00:25:04] Andrew Zigler: And so really interesting deep dive. Loved the opportunity to share some of the anecdotes. I guess, unfortunately, this is another factory. So how many factories have we talked about now? So this is a content factory. We talked about the twilight factory. We talked about the dark factory. If you're interested in how this kind of a process does apply for content, uh, definitely give this one a read, and if you have thoughts, reach out.
[00:25:29] Andrew Zigler: We'd love to hear, uh, your take on our agentic newsroom or what ideas you have to make it better.
[00:25:34] Adam: yeah, and o- obviously this story's, um, you know, really fun from a, personal perspective. It's, it's fun to get outside reporting, um, on, on the stuff that we do here on Dev Interrupted, and I think is also, very rightfully highlights you and how much you've orchestrated behind the scenes to help the show run.
[00:25:55] Adam: And, and you've built so much cool tooling in, in a way that's I was actually [00:26:00] thinking about this, that's almost been, been bad for my own personal AI development. And, and what I, what I mean is, like, it has been so easy for me to rely on some of the great stuff that you build internally. So, you know, normally you and Ben end the segment by, by talking about what your agents are building this week.
[00:26:19] Adam: And so I actually, for the first time ever, have my own harness. It's up and running. I'm using it. Um, but I feel like I'm behind the, the curve a little bit, at least on the bleeding edge as, as I do think you have continually been in your own, AI enablement journey. And so I'm, I'm, I'm at, I'm at the start now of that journey on my own.
[00:26:45] Andrew Zigler: you're at the, you're at, you're at, you're at your own starting line. Maybe, okay, so, so what I'm hearing is, is that I had the opposite effect of, of an agentic halo. I was an agentic, like, event horizon, an agentic, like, black hole because then you end, you [00:27:00] ended up utilizing so many of the things that I was building instead of building those things yourself.
[00:27:05] Andrew Zigler: so that's definitely uh, like an opportunity I think even for me to help you, you know, get more agentic and maybe that goes for all of our listeners too 'cause we talk a lot about, you know, my factory and the things that I work with and my agents and I think there's a lot of opportunities to show it as much as we do tell it which is why I'm really excited about how we're gonna be evolving Dev Interrupted.
[00:27:27] Andrew Zigler: You know, this, this article is a glimpse inside of our agentic newsroom as it is now and it's in the process of evolving. What I've learned is that there's so much in this factory that we've built that I can show y'all and also so much that's more demonstrable than just, like, being in a podcast and so we're gonna be exploring things like doing a live stream and doing, uh, coding sessions as well because we, every week we cover these really interesting tools, these really, really useful ways of working with harnesses and agents, uh, to achieve the work that you do every day [00:28:00] whether it's like you're a developer or you're an engineering leader and you're leading a team.
[00:28:03] Andrew Zigler: We're all expected to have a level of fluency with these tools so I'm gonna be excited to start showing some of that, in the future because on that same note we're taking a break next week, or rather on the new segment because I'm gonna be in New York. I'll be presenting at LDX3 NYC which is Lead Dev's flagship conference in New York City. I'm gonna be giving a talk about why Miss Frizzle would be a great SRE and if you've been listening to this podcast for any amount of time you could imagine what that might entail. And so if you're in the New York area and you're at Lead Dev I'd love to hear from you. Be sure to stop by. But, um, after that week we'll be back with a, a newly envisioned new segment that's going to be, uh, terminal first.
[00:28:45] Andrew Zigler: I'm gonna be bringing you along and helping you develop the skills and the harnesses and Adam maybe this is an opportunity for you to maybe learn alongside some of our listeners too and we can figure out how to make, uh, everybody more successful with agents and, and [00:29:00] not just have the a- Dev Interrupted agentic newsroom be one, you know?
[00:29:04] Adam: Yeah, 100%. I, I think too you, you also raise an, an interesting point about kind of like the shape of what it would look like in a team. Because, you know, you talked about, like we've talked a ton about, right? Like you're a 10x developer or a 20x or 100x, right? But, but how, how does that work in relation to the team around you, right?
[00:29:23] Adam: How does it work to people to your left, to your right? Like maybe you're faster, but y- the whole team's not really faster. To that point, I don't know that it's a bad thing, if people on a team don't have a great grasp of the tools. Because by extension of being underneath the umbrella you built, I was already incredibly faster.
[00:29:46] Adam: And so yes, I have my own interest in wanting to learn these tools and, just professionally grow, my ability. But I think there's also [00:30:00] room for somebody on a team to be like, "I don't want to know. I don't care to know." And I suspect that that's probably gonna happen, right, at a whole lot of companies.
[00:30:08] Adam: And
[00:30:09] Andrew Zigler: agree
[00:30:10] Adam: if, there's enough people like you, like it kind of won't matter as long as like it's pulling the entire team with them. And in this case, for Dev Interrupted, that's, that is, is exactly how it's worked, which is I think probably like a good understanding of like, yes, people should have a baseline understanding of how these tools work.
[00:30:30] Adam: But like as more industries adopt these practices, I think there's gonna be a whole lot of versions of me and kind of my experience, but then that don't decide like, "Hey, I wanna build the harness too," right? Like they'll probably just be like, "Cool, I'm a whole lot better at my job now."
[00:30:49] Andrew Zigler: Yeah. And you know, you make a really interesting point that it's like everyone's gonna fall on different levels of the gradient of, like, how much of it they're gonna be involved with in building themselves. The challenge that I will throw back at [00:31:00] you is that I do think that, you know, you describe this world where you have some, like, this, like, cohort of agentic things that are just, like, holding up the entire ether of the company, and I actually don't realistically think that'll be the shape that these kinds of things will take.
[00:31:16] Andrew Zigler: I think that because that kind of group or that small of an amount of individual can hold up so much, you're gonna see a lot smaller companies and you're not gonna see these other decision makers that are outside of that system. Because that's a core part of, like, the factory analogy and about, like, the decision managers, the decision makers.
[00:31:35] Andrew Zigler: Like, know, at the beginning of this year, I made a joke about Gas Town where I was like, "Oh," like, "what's gonna happen?" I was like, "Your Gas Town's gonna call my Gas Town and they're gonna
[00:31:45] Adam: Yep.
[00:31:46] Andrew Zigler: And, like,
[00:31:47] Adam: Yep
[00:31:47] Andrew Zigler: fun of it, but, like, I actually fully do think that that's the shape that things go towards.
[00:31:53] Andrew Zigler: Because, you know, this agentic newsroom I've built, these factories that people have built, engineers have their [00:32:00] own mini version of that factory themselves for any little thing that they ... For the thing that they specialize in doing. The factory isn't really something that exists outside of the developer. And so when you have a company and the idea is that the company is the factory, for me what that means is that the company is synonymous with a small group of developers who are all micro factories working together or are coordinating on one larger factory. But that would be the shape of the organization.
[00:32:30] Andrew Zigler: You know, everybody that's plugged into that would also be conduiting something into it. There wouldn't be those outside of that token sphere, I guess you could say. And this sounds really wild and I think it's, like, a far future world of, like, where this might go, but I do think that in general There's gonna be a huge incentive for no matter what your skill or your trade is or where you work, for you to be thinking about what is my own factory?
[00:32:54] Andrew Zigler: What's my own thing that I build and my own that I take from job to job, that I [00:33:00] develop over time, that I fine-tune? 'Cause, uh, it's kind of like being a chef and, like, taking your knives to, like, your job. Like, right, you own those
[00:33:08] Adam: Yeah, I like that analogy
[00:33:09] Andrew Zigler: are my cooking knives, right? It's like that's actually how engineers are building this.
[00:33:13] Andrew Zigler: These factories that we're building, like, yes, like, companies own them in name and they represent companies, but at the end of the day, these factories are just outward projections of the engineers that are extremely agentic, right? And so, uh, I think it's really gonna be interesting to see how it evolves.
[00:33:31] Andrew Zigler: That's why I'm also really interested to bring y'all along in the terminal to show you my factory and how I've thought about it, and maybe give you some ideas for starting your own. Um, and you know, the last thing I'll say here, so a really cool thing happened this week from our friends at AWS, uh, that I really just can't go through this news segment without mentioning. And, you know, we've talked about AWS and specifically Kiro here on Dev Interrupted many times. We've had, we've had a few Kiro, uh, leaders here on the show. I even was on a Kiro [00:34:00] live stream, a few months ago, and I, I love building in Kiro. We've, we've been talking about Kiro since it was provided by AWS.
[00:34:07] Andrew Zigler: And, uh, a really amazing development came out this week where they are providing Kiro Pro, uh, to 132 universities, where you're talking about, like, over a million college students that now get access to a free year of Kiro Pro on AWS. That's 1,000 credits a month. That's a lot of API tokens and consumption that you could use to build stuff.
[00:34:27] Andrew Zigler: And so if you're, if you're in college or if you have, you know, someone that you know that's a university student, they don't even have to be computer science. They should reach out, go to the kiro.dev/students website and get their, uh, plan, and be thinking about how maybe to turn, uh, that opportunity into whatever factory they're gonna be building and owning in the future.
[00:34:47] Andrew Zigler: It's a great opportunity, um, and there was a amazing video that came out of AWS
[00:34:54] Adam: So good
[00:34:55] Andrew Zigler: us cracking up. So if y'all, if this is, if this is [00:35:00] news to you, go, go to LinkedIn. Come find me on LinkedIn. Check out the video that I shared as a response to the one AWS sent us. It's a total crack-up, and you get to see me be, like, a total goofball.
[00:35:11] Andrew Zigler: So, uh, this is only one part of that story. Be sure to go check out the rest of it.
[00:35:16] Adam: Yeah. Plus, plus one to that. Phenomenal video. That was so fun watching them send that over to the team. and then I think the response video that we worked on is equally fun. So yeah, go, go check out both of those. big fans of what they're doing over there. All right. I think that is it for this week's edition of the Friday Deploy.
[00:35:39] Adam: It was fun. It was fun to actually
[00:35:41] Andrew Zigler: Great having you
[00:35:42] Adam: step forward. I know
[00:35:43] Andrew Zigler: Yeah. Great pulling, pulling you out from backstage and having you up here. Maybe we'll have to do it again sometime. Uh, but I'm real excited, uh, to be taking y'all into the terminal in the near future
[00:35:54] Adam: Yeah, I'm really excited for it too. All right. Thanks for having me on.
[00:35:58] Andrew Zigler: See you all next time.
