Sign in or Join the community to continue
Bridging the Gap Between AI Builders and Domain Experts
Posted Oct 12, 2026 | Views 12
Share
Speakers
Justin Schroeder
Co-founder @ Standard Agents
Justin Schroeder is the co-founder of Standard Agents and creator of open-source developer tools including FormKit, AutoAnimate, Tempo, ArrowJS, and dmux. His work centers on giving complicated software problems clearer boundaries: first in web interfaces, and now in systems that coordinate specialized AI agents.
+ Read More
Demetrios Brinkmann
Chief Happiness Engineer @ MLOps Community
At the moment Demetrios is immersing himself in Machine Learning by interviewing experts from around the world in the weekly MLOps.community meetups. Demetrios is constantly learning and engaging in new activities to get uncomfortable and learn from his mistakes. He tries to bring creativity into every aspect of his life, whether that be analyzing the best paths forward, overcoming obstacles, or building lego houses with his daughter.
+ Read More
SUMMARY
Everyone is racing to build one agent that can do everything. Justin Schroeder, co-founder of Standard Agents, thinks the future looks more like thousands of tiny ones. Ahead of his talk on domain-specific agents at AGNTCon + MCPCon North America, he explains why tightly scoped agents might be the smarter bet, even if it feels like a step back toward the single-purpose machine learning models of a few years ago.
+ Read More
CONTENT & TRANSCRIPT
Justin Schroeder: [00:00:00] There's going to be a future where our agents are tuned to perform one small task and perform it really, really, really well. And in that case, you could actually sort of measure, uh, the efficacy of how good it can do whatever that thing is.
Demetrios: That's literally my buddy Raul just told me like, "Oh man, I left my stuff running and I got a 1K bill I was not expecting."
Demetrios: Like software factories are hard, man. If you throw something at a big model and it does it, like is there really the need to go and then scope down like you're talking about? If you've got these thousand agents, isn't that so taxing on the human brain to figure out like where's that like issues agent? I can barely keep track of my freaking chats that I have right now and I don't have these thousand agents.
Justin Schroeder: It's a good question because of, of two things. One[00:01:00]
Demetrios: All right, Justin. I'm excited for your talk at Agent Con and MCPCon in October in San Jose. This is the North American talk. What are you gonna be talking about?
Justin Schroeder: So the title of the talk is "The Future is Domain-Specific Agents." Mm-hmm. Um, and it's basically that. There's this sort of, uh, rustling in, in the, in the trees right now about this thing called domain-specific agents, although everybody has a different term for it.
Justin Schroeder: So Jensen- Yeah ... calls it, um, super agents. Jensen calls it super agents. Uh, Elon has called it, like, specialist agents. Um, but the idea is that there's a huge unlock where you have small agents that are around a very narrow set of tasks. Hmm. So where our models right now are being trained to be sort of God models that have, you know, they know how to code, but they also know how to, like, create a [00:02:00] water filter to survive an, an atomic apocalypse for some reason.
Justin Schroeder: Like, they can do both of these things. I see. There's going to be a future where our agents are tuned to perform one small task and perform it really, really, really well. Um, and, and in that case, you could actually sort of measure, uh, the efficacy of how good it can do whatever that thing is, however small it is.
Justin Schroeder: Like, maybe the entire agent's job is to, like, check Gmail.
Demetrios: Mm-hmm.
Justin Schroeder: And that's all it can do. So it's even more fine-tuned than what you'd think of, like, a Grok bot or something like that. Um, and if you could do that, if you could have a little agent that was just like, it was perfect at checking Gmail, like, that's all it has to do.
Justin Schroeder: How hard can that be? If you could do that, then maybe you would wanna share that agent with other people Right? And you could post it online i- in, you could have it on a thumb drive, you could have it however you want, and then that agent could work with another agent, and these things could compose a little bit like we do with skills right now, but with agents [00:03:00] instead.
Justin Schroeder: And the only reason that this doesn't exist is, well, first of all, uh, it's sort of a higher, it's sort of a higher order, uh, thought process that we haven't quite, you know, like we're unpacking this as we're all learning together. Um, but the only reason it really doesn't exist is that there's no standards around it.
Justin Schroeder: Mm. So if I create that incredible Gmail agent and I give it to you, you can't use it. So- Makes sense ... we have to solve the standards problem. So my talk's gonna be a lot about that.
Demetrios: Oof. Oof. All right. So thousands of agents that are hyper-scoped in a way.
Justin Schroeder: Yes.
Demetrios: Or thousands of agents that have very, very tightly scoped tasks that they are trying to do.
Demetrios: It feels like we are going a little bit back in time and just like machine learning models used to do that, right?
Justin Schroeder: Yes, machine learning models did used to do that. The cool thing now is that you don't have to train the model. So- [00:04:00] Yeah ... that, uh, the GEPA paper or GEPA, however you choose to pronounce this word, was really instrumental in, in setting the stage for some people, but it basically says that it is more effective to train your prompt than it is to train your model.
Justin Schroeder: Uh-huh. And there's a really weird relationship between a model and a prompt and the tools that are given to it that is not very well explored right now. Um, and you've seen the labs hint at this from time to time when, like for example, uh, OpenAI recently said, "Hey, actually, uh, it turns out that our model performs way better on this test if it's not using this old broken down harness."
Justin Schroeder: "If it's using this harness- Mm-hmm ... then it's way better." Well, that's actually true for literally everything. Everything is like that. Um, you can have an unbelievable output from one harness and a terrible output on [00:05:00] the same prompt with the exact same model, and the only difference is the harness. Now, I've got tons of examples of this.
Justin Schroeder: So what you're searching for is to find effectiveness. You want your agent to be very effective, and once you've locked in that effectiveness, it almost doesn't matter what model it's using or how old it is or something like that. Mm-hmm. So our coding agents right now, everybody is being trained that we just like, oh, we see a new model release and we just switch our coding agent to the latest model and boom, we're good.
Justin Schroeder: Um, that's actually fine at these like really large scale models. It like kind of can figure it out, although you are losing something in there. Um, but with small models, this is, this is not possible. So y- we really need a type of specification that says this exact prompt with this exact model hosted on this exact provider, you know, this entire chain of things, uh, has provable efficacy for doing the task that [00:06:00] it's, that it's made to do.
Justin Schroeder: And then when- Mm ... I share that with you, you're getting the exact same efficacy.
Demetrios: Ooh. And so you see standards as almost being like metadata that you can pin to an agent so that it knows everything that it needs in its harness to be effective?
Justin Schroeder: Yeah, not only that, I think that the, I think that there, there really should be a runtime for agents.
Justin Schroeder: Mm-hmm. So imagine, uh, in the same way that we have a runtime for Node, and Node can run in Bun or it can run in, uh, Deno or it can run in Node, Node, uh Uh, uh, QuickJS. There's lots of different runtimes that you can run JavaScript in or the browser. Um, we need runtimes for agents that's very similar. But all of those languages, or all of those runtimes are, are complying with the same standard specification.
Justin Schroeder: Right. So there's gonna be variations in them. The V8 runtime [00:07:00] is just really fast, you know, and it's faster than other ones. So you might wanna use that runtime to run your, you know, mission critical application, but that's not the only thing that matters. And so I could see an ecosystem of runtimes that exist that are all able to run your agents in a guaranteed way, um, that have different pros and cons.
Justin Schroeder: But w- that is the world that I am 100% convinced is going to emerge. Um, people don't believe this right now, but I'm 100% convinced that eight months from now, this will be a no-brainer for everyone.
Demetrios: Oh, dude, I'm so glad you just said that 'cause I already see the clip forming in my mind. Just as a li- of just like, uh, we start with that and then we go like, "Elon calls it this, and Jensen calls it that," and then boom, like I, I have the voice overlay of Justin is bringing ba, ba, and then I talk- Boom
Demetrios: about what you're doing. And then we rock and roll with a few more clips of like, "But we need standards, and that's the [00:08:00] key here." Uh, and then there's your little trailer that we've got. So note to editor, that's how I'm seeing it in my head. Uh, and, uh, all right. But there is, there's a few cool things that I just wanna like satisfy my own itch for when I ta- when we talk about this.
Demetrios: I've found that you get to a certain place and you need this out of, or you, you take this step of going very, very small scope with small models or smaller, like, scoped models. I- it doesn't necessarily have to be a small model. When you have like a forcing function, right? It's not necessarily, you don't necessarily think like when you wake up in the morning, you're like, "Wow, cool, let me go and make 1,000 super small agents that just do one thing really, really well."
Justin Schroeder: Mm. Yeah.
Demetrios: And you take like a microservices paradigm to your [00:09:00] agent building because If you throw something at a big model and it does it, like, is there really the need to go and then scope down like you're talking
Justin Schroeder: about? Oh, this is a, this is a great question too. Um, there must be, because on multiple vectors.
Justin Schroeder: On one, there's the efficacy itself, and you're right that a big model is generally gonna have better efficacy than a small model on any given domain. But that's not true on any given domain if, if you've had, like, a really reinforced, uh, prompt and model. And like, for example, a, uh, my, my coding agents have, like, I don't know, half a dozen to a dozen skills on them, a few MCP servers.
Justin Schroeder: And so every time it needs to do work, it's sort of wading through this ocean of stuff, and it has- Yeah ... tons and tons of different tools. But if I were to create that Gmail agent, it [00:10:00] would have Gmail-related tools. Like, it would have a tool to be like, "Check latest messages," right? And like, that's all it can do.
Justin Schroeder: And so you actually can squeeze out way more value from a smaller model very quickly, and this is actually why I don't think routing is going to end up being the answer. I don't think model routing will end up being the answer, because even the big models don't really know how hard a problem is on the front side.
Justin Schroeder: You know how hard a problem is on the front side, because you are faced with it, just like you said. At some moment, I'm, I, like, have a need. I need this thing to, like, do a background job. I just, like, I just need it to go check issues, uh, on a five-minute loop, and then notify me or queue up something else.
Justin Schroeder: That's all I need. And it turns out I just burned 60% of my Fable weekly allotment on that exact problem this week.
Demetrios: Dude, that's what literally my buddy Raul just told me, like, "Oh man, I left [00:11:00] my stuff running and I got a 1K bill I was not expecting." Like, software factories are hard, man.
Justin Schroeder: Dude, they are.
Demetrios: Yeah, especially you, it can get away from you quickly.
Justin Schroeder: So imagine if, like, when I hit that exact problem, then I tackled it. I was like, "Oh my gosh, this is ridiculous." There's so many advantages if there's standards. First of all, when I say, "Okay, I need to build that agent," my agent knows exactly how to build it.
Demetrios: Hmm.
Justin Schroeder: I mean, what I'm not saying here is like I'm, I have the spec and I have the runtime, so that exists.
Justin Schroeder: Uh- Yeah ... which is why, like, I get to play with these things all the time and I can see that they're very effective. Uh-huh. Uh, but, but I can just spin up any, any of my coding agent, Fable or, or Codex or whatever, and I just say like, "Hey, I need you to build an agent that does, like, this discrete task," and then it just goes and does the discrete task.
Justin Schroeder: Uh, it creates the agent with the precise tools and everything, and now it's, it, I can, I can deploy it on the cloud, I can give it to my coworkers. It's, it's like a very, uh, distilled thing. It, I, I'm [00:12:00] specifying absolutely everything down to the model. There's all kinds of problems that come with this, um, that you have to solve, which is why it's sort of an unsolved problem at the moment.
Justin Schroeder: Um, but it will get solved.
Demetrios: I think I've heard this called agents as tools. Would that be the same thing?
Justin Schroeder: Yes, it could, it could be. Um, that, in fact, that's like a way to solve the problem.
Demetrios: Yeah, 'cause what I'm understanding from you is more like, okay, there's the... A- and you're saying specifically maybe it's even not agents as tools Because the model router of the, like, big brain that goes and chooses and kicks off tasks to sub-agents isn't necessarily the best way of going about this.
Demetrios: Especially if you already know, "Hey, I just want this thing to run, uh- Yeah ... every five minutes." And it's not like it's this very complex thing, and I can hyper-optimize this thing that just needs to happen. It grabs issues or it tells me what my [00:13:00] latest unread m- messages are. You know? Like, those things can happen on a very frequent cadence- Yeah
Demetrios: and you don't need a big model for them, and you know that it needs to happen, so-
Justin Schroeder: That's
Demetrios: right ... you just set up, you set up this specific agent to do that. But where I am kind of still questioning is how this looks in practice. If you've got these thousand agents, isn't that so taxing on the human brain to figure out, like, where's that, like, issues agent?
Demetrios: I can barely keep track of my freaking chats that I have right now, and I don't have these thousand agents.
Justin Schroeder: Such a good question. Uh, it's, it's a good question because of, of two things. One is, like, this distinction between sort of a coordinated with sub-agent structure, like a tree structure of agents or, or graph- Yeah
Justin Schroeder: engineering is kind of what people are calling it now, right? But, but basically a tree of agents where you have some sort of a coordinator sitting above a bunch of other ones. Um, that's a great architecture [00:14:00] for all kinds of things. So if you were to take this same idea and try to build, like, a Hermes, OpenClaw style personal assistant, that's what it would look like.
Justin Schroeder: You would have, like, one that you're speaking to. It has a lot of sub-agents. They're all going and doing their different tasks. It would look like that. But that is by no means the only mechanism by which agents are going to find penetration. So yes, it can do that, but also it can just be this, like, tiny little back-of-house thing that's running for a butcher shop, and it's, like, figuring out how many pounds of meat to, uh, order for the next week.
Justin Schroeder: That is, is, like, super realistic, like a very realistic, tiny little agent that would cost them absolutely nothing to run every week. And this is like an advantage that I actually have not living in the Valley right now Mm-hmm. Is, you know, every- when I go to the Valley pretty frequently, and when I'm there it's like, "Oh my gosh, this is amazing.
Justin Schroeder: There's so much hype and excitement." And then I like come out, [00:15:00] out to... I'm in Virginia, I come out to Virginia, and it's like nothing. Nobody, zero penetration.
Demetrios: Crickets. Yeah.
Justin Schroeder: Absolute crickets. I could walk around and be like, "What are you guys doing with a- with agents these days?" Nothing.
Demetrios: No.
Justin Schroeder: Yeah. Yeah.
Justin Schroeder: Absolutely nothing. What are you, what are
Demetrios: you talking about? ChatGPT?
Justin Schroeder: That's it. Yeah. Yeah, exactly. Yeah. But like this is, this is sort of, in my opinion, the missing link. Like it is the thing that will m- that will turn this into, um, a Main Street tool, is being able to say like, "Actually, we've got this, this very narrow implementation that meets your needs precisely-" Mm.
Justin Schroeder: uh, rather than a $200 Claude code sub or something."
Demetrios: You know where I have heard, and I actually was just thinking about how strong of an opinion I have on this, that it won't happen. So potentially we can talk about why- Yeah, yeah, yeah ... you think it will happen right now, is since 2022 or whenever agents started getting popular-
Justin Schroeder: Mm-hmm
Demetrios: the first [00:16:00] thing that a lot of people went to try and build was an agent marketplace. And I am- Yes ... convinced that is never going to happen. It's like "Mean Girls," like stop trying to make fetch a thing because it is not gonna be a thing, in that, again, like me as a person, I'm not gonna go sift through a bunch of agents that random people built on the internet and see if it might potentially work for my use case, but then I gotta tweak a few things or I gotta tell Claude to tweak a few things.
Demetrios: Instead of just, why don't I just tell Claude to start it from scratch and exactly what I need, and I create that agent, and then if I can distill it down and make it a smaller agent like you're talking about, I see the value in that. But an agent marketplace, I'm 99% sure that's never going to be a thing.
Justin Schroeder: I think it's already a thing.
Demetrios: All right. So good. You, you got the counter. Hit
Justin Schroeder: me with it. I think it's already a thing. Everyone just has their own marketplace. Um, and the only [00:17:00] place that we've shared any sort of agreement on this right now is MCP and Skills. But MCP and Skills are effectively just like really bad agents, 'cause they're just trying to convert the agent that you're using into one that can do these other things, which has this like incredible upside, which I'm not gonna try to take away from it, which is, uh, that now I get to still use the same user interface or something like that.
Demetrios: Yeah.
Justin Schroeder: The downside is that it's all context injection. It's like a giant rag pipeline just getting shoved into the brain of, of this agent. And like, not only do the agents get worse when you do this, um, they, they get dumber on like completely related tasks. So like if you were to take that same model and the same agent, like, like Sol, and give it the job of a skill exclusively, it will do it way better than if you have that skill get injected on top [00:18:00] of its preexisting everything.
Justin Schroeder: Mm. Right? So the argument here is not that like people are gonna need to go out and like assemble these things carefully themselves. They're going to use them very much in the same way that people use skills right now. Um, which you are, you probably have already done this. Somebody somewhere in your life has been like, "Dude, you gotta check out a skill.
Justin Schroeder: I know a skill, skill, whatever." But like, you know, so you
Demetrios: got- Superpowers. Shout out, shout out to superpowers. There you go. Exactly. It's always superpowers or, or the Grill Me, or last week it was Show Me. It's true. I have downloaded a whole lot of- You've done it ... skills from- You've already done the thing you said you're never gonna do.
Demetrios: Yeah, from GitHub. I, I did. That is a marketplace, you're right. So that's-
Justin Schroeder: It is a marketplace. But imagine if what you were installing was like actually effective agents. So there's no reason why the interface for domain-specific agents couldn't be your existing agent. There's no reason why that's not the way it's exposed.
Justin Schroeder: It could be exposed that way. Um, but it just needs its own guaranteed way to execute [00:19:00] and run and so on and so forth. So like that, that could be one, one output. But I'm not even really interested in like what that final interface is. Hmm. I'm interested in trying to say we need a way to share and distribute these things, and like that might be the, the, the winning outcome.
Justin Schroeder: Maybe the winning outcome is every human on earth has their own website with its own chatbot on it. That sounds very unlikely, but I don't know. Or, uh, you know, maybe GrokBot ends up becoming like the thing and everyone- Hopefully not ... has a GrokBot.
Demetrios: Yeah. '
Justin Schroeder: Cause they're doing it too, right? Like they have their- Yeah
Justin Schroeder: own marketplace. They, if you've used GrokBot, they have their own marketplace. You log in there, you find the, their plugins, like you wanna use Google, it's gotta be one of their plugins. And it, it's like all... I mean, most of them are MCP related, but like it's still this like pre-approved marketplace of things that plug into GrokBot.
Demetrios: Actually, as soon as I created my Hermes agent, the first thing I did was like go out and scour the internet for [00:20:00] what are the best Hermes use cases and skills that I should have with my agent. Yeah. And so if you look at it like that- So yeah.
Justin Schroeder: Oh, man ...
Demetrios: then it's true.
Justin Schroeder: The use ca- I know, like we're already over time, but the use case thing is actually one of the most interesting problems too, is like when you go and you talk to people about like, "Hey, what would you use a personal agent for?"
Justin Schroeder: Yeah. Uh, if, if somebody's not using it already, or even if they are- Yeah ... using it. They're like Hmm. Calendar maybe, like a- Yeah. ... maybe my email? That's the thing.
Demetrios: I had to do some digging. It, it
Justin Schroeder: was
Demetrios: exactly that.
Justin Schroeder: This is, this is why domain specific agents are the win, because people in different domains know how they would use it in their domain.
Justin Schroeder: Hmm. Right? Like, when you're at work, and if you're not working on agents, which is sort of this vicious, you know, what is it? A, something eating its own tail. Mm-hmm. A self-licking ice cream cone. Um- [00:21:00] That's... Yeah ... if you're not doing that, and you're, like, a nurse or you're, you know, some sort of doctor, like, you immediately are like, "Oh my gosh, I have so many use cases where, like, I could have it, you know, be doing the, the transcription or the notes, or it could be doing..."
Justin Schroeder: Like, there's so many different use cases that we in the tech world are just, like, sort of immune from. Um, but the domain experts need to make the domain specific agents. So they need a way to just do that.
Demetrios: Yeah. They need a way to do that, and then sh- I think where I've seen it fall down in the past is that they don't realize what it, what the possible is.
Demetrios: And so- Yeah ... like the transcription, for example, it's like, "Oh, it can do that? Oh, well yeah, like I would love to get transcribed- That's true ... all of these different things that happen," right? And then you start, you start to unlock these doors and you realize, "Wow, that's, that's, uh, that's possible? Okay. Well, yeah, then I want that agent too."
Demetrios: Yeah. And so then there's this merging of the two worlds, of the [00:22:00] tech, where we know everything it can do, but we don't necessarily have the use cases, and the- Yeah ... folks that have the use cases who wanna know, like, what can it do?
Justin Schroeder: Yeah. This, this h- we need to figure out a way to empower that.
Demetrios: Well, I'm so excited for your talk.
Demetrios: You've got me convinced. I'm gonna be there, enjoying it, and hearing about this whole world of domain specific agents, and I'll see you October 22nd and 23rd in San Jose for AgentCon and NCPCon. I'll be there.
Justin Schroeder: Looking
Demetrios: forward to meeting you- Thanks for being
Justin Schroeder: here ... and everyone else
+ Read More
Watch More
Bridging the Gap Between AI and Business Data
Posted Jun 20, 2025 | Views 193
# AI and Business Data
# LLM
# Snow Leopard AI
Bridging the Gap between Model Development and AI Infrastructure // Mohan Atreya // AI in Production 2025
Posted Mar 13, 2025 | Views 360
# LLM
# GPU
# Rafay
PyTorch: Bridging AI Research and Production
Posted Nov 15, 2021 | Views 349
# PyTorch
# PyTorch Ecosystem
# AI Models
# Fireworks.ai