AI Agents Are Here: What They Can Already Do—and What’s Next (Stanislas Polu & Harrison Chase)

What’s next for AI agents, and how will they change the way we work? In this conversation, Stanislas Polu (CEO of Dust, formerly research at OpenAI) and Harrison Chase (CEO of LangChain, one of the most influential open-source AI frameworks) unpack the current state and future of AI agents. They reflect on their early conversations in the pre-ChatGPT days, how the landscape has evolved, and where it's headed next. Stan and Harrison share lessons from building today’s agent infrastructure—from chat interfaces to the future of ambient, autonomous systems—and discuss the challenges of operating in the chaotic "fog of AI." We dig into the open questions, early insights, and messy realities of building in today’s fast-moving AI landscape. In our conversation, we explore: • What sparked Stan and Harrison’s early interest in LLMs • The pre-ChatGPT era and how the AI landscape has evolved since late 2022 • High-leverage use cases for agents inside Dust and LangChain today • The critical differences between AI workflows and true agents—and why agents may unlock more powerful, long-term solutions • Why reliability is the main blocker to ambient agents • Real-world enterprise use cases for AI agents across customer support, sales, and engineering • How to build in the “fog of AI” and the challenge of maintaining product vision when foundations shift every six months • Strategies for creating defensibility in a world where tech giants can quickly replicate features • The future of multi-agent systems and how they could transform enterprise productivity • The current state of the AI talent market • And much more — Thank you to our sponsor: Brex https://www.brex.com/mario—The banking solution for startups. — Transcript: https://www.generalist.com/p/the-evolution-of-ai-agents — Timestamps (00:00) Intro (02:33) Brief overviews of Dust and LangChain (03:30) The early days of LLM product development (11:02) Harrison's journey to founding LangChain (14:35) Dust's evolution and focus on enterprise productivity (17:15) Tobi’s AI memo (18:42) An overview of AI agents and how they differ from AI workflows (26:43) High-leverage use cases for agents at Dust and LangChain (30:41) How to interact with agents and an explanation of ambient agents (36:21) What the future of agents may look like (40:52) Current limitations of AI agents and reliability challenges (45:40) Will we converge to one agent or many specialized ones? (51:32) How to solve the sycophant problem (56:04) The challenges of building AI companies in a rapidly changing landscape (01:03:06) Recent AI talent acquisitions and market dynamics (01:05:28) How Dust and LangChain attract talent (01:09:12) How far off AGI may be (01:12:52) Final meditations — Follow Stanislas Polu LinkedIn: https://www.linkedin.com/in/spolu/ X: https://x.com/spolu — Follow Harrison Chase LinkedIn: https://www.linkedin.com/in/harrison-chase-961287118/ X: https://x.com/hwchase17 — Resources and episode mentions: https://www.generalist.com/p/the-evolution-of-ai-agents — Production and marketing by penname.co https://penname.co/. For inquiries about sponsoring the podcast, email [email protected].

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  The big question we see in the market today, which is interesting, is agents versus AI workflows. We do foresee a world where those agents will be actual coworkers. And I don't think you can really encode a coworker with a workflow. We're really bullish on trying to help people create agents, not workflows.

• 0:15Mario Gabriele?

  What is the right way to think about what an AI agent truly is and what it isn't?

• 0:19Speaker 02

  You can often do the same things with workflows and agents. It's just the ease of how you describe it. In an agent, it would all be in natural language, right? Like, you could have a recipe that you just put in natural language. You say, hey, do A, then do B, then do C. And it's not as deterministic, so it's not as sick, but it's way easier.

• 0:34Mario Gabriele?

  I'd love to zoom out for a moment and talk also just about what it means and what it's like to be building in AI at the moment and some of the specific dynamics that founders have to face.

• 0:44Speaker 00

  The fog of AI is the fact that the foundations are moving very quickly. And so you have to have a vision of where you're going. But you cannot paint it because if you paint beyond six months, whatever you wear, painting will probably be not true.

• 1:01Mario Gabriele?

  Hey, I'm Mario and this is the Generalist Podcast. As the saying goes, the future is already here. It's just not evenly distributed. Each week, we sit down with the founders, investors, and visionaries living in these pockets of the future to help you see what's coming next. Today, I'm speaking with Stan Polu and Harrison Chase about the future of AI agents. Stan is one of the founders of Dust, a Sequoia-backed platform that makes it easy for enterprises to build and deploy custom agents across their workforce. Before that, Stan was at OpenAI, researching the mathematical reasoning capabilities of large language models. It was during his time at OpenAI that he first met Harrison Chase, who went on to found LangChain, a popular developer framework

• 1:43Mario Gabriele?

  that has become essential infrastructure for building agentic applications. LangChain has also become one of Silicon Valley's hottest companies, raising from Benchmark, Sequoia, and IVP. Both Stan and Harrison spend their days thinking deeply about AI in general and agents in particular. In our conversation, we explore what comes after chat, why there probably won't be one agent to rule them all, and the unique challenges of building a business in the fog of AI.

• 2:11Mario Gabriele?

  I walked away from our conversation with practical insights about where agents are headed and how they might change the way we work, think, and build over the coming decades. This is a new podcast, so if you enjoy today's episode, I hope you'll consider subscribing and joining us for the incredible conversations we have coming up. Now, here's my conversation with Stan and Harrison.

• 2:33Mario Gabriele?

  Thank you both so much for for being with me today. I'm excited to to chat about AI in general and sort of agents in particular with two people who are really spending their lives dedicated to this, you know, field and and this part of the world. Maybe to start, we could just begin with a little bit about both of your your companies and what they do. So Stan, why don't you tell us about Dust?

• 2:54Speaker 00

  Yeah, well, Dust is the place where you build agents for for Walker. So it's a product that lets you create, manage, and operate agents with access to your company context and tools.

• 3:06Mario Gabriele?

  Amazing. And we'll get into the the details of what that means and and why it's become so powerful more, I'm sure, over the course of of this next hour or so. Harrison, yeah, I'd love to to have folks know about Langchain.

• 3:19Speaker 02

  Absolutely. At Langchain, we build developer tools to try to make it as easy as possible to build agentic applications. That's the one liner, but I'm sure we'll get deeper into it over the course of the conversation.

• 3:31Mario Gabriele?

  Amazing. Well, I think something that you both have in common is that you've been in AI, at least by the industry standards for a long time, you know, predating sort of the ChatGPT moment in in November of 2022. I started in VC in 2016 when I was my first venture job. And I remember there was a big sort of excitement around those first phase of chatbots. But, you know, honestly, lots of the excitement sort of withered for the years that followed. And I think there was a question mark about when exactly AI and the way that we see it today was really going to come to the fore. And so I'm curious for you both, you know, what was it that you were seeing in the industry before it became so obvious to everyone else when ChatGPT came out that excited you to sort of like build your career around this? And and maybe let's start with you, Stan, because I

• 4:16Mario Gabriele?

  know you spent some time at OpenAI, so you probably had, you know, a front row seat to a lot of of the These questions.

• 4:21Speaker 00

  So actually, the fun fact is that we chat together with Harrison. I think it was maybe September twenty twenty two or maybe October twenty twenty two. It was funny because it was a pre-chat G P T and we we naturally met together and spent some time chatting together because there were so few people like interesting themselves in the use of L L M's for development or use of L L M's for doing anything else in product. And so that's for the the fun fact to to to give everybody the kind of a context of how small the community was during those few months that were the end of the summer twenty twenty two, all the way to chat G P T release, which was in October or, yeah, I guess, October.

• 5:04Speaker 00

  So it's a fun fact. As far as I'm concerned, I had the the chance to work at OpenAI, so I had been working on the studying the mathematical reasoning capabilities of L L M's, but obviously was also exposed with G P T four, which had under training early summer twenty twenty two. I mean, it was pretty obvious to me that there was a massive disconnect between the capability of the technology and the actual impact it was having on the world. The revenue of of OpenAI, probably not public, and I don't even remember the right numbers, but it was like a drop compared to what it is today. There was just a few tens of millions of dollars. And so compared to the power of G P T four, there had been the luck of playing with internally.

• 5:46Speaker 00

  I think it was obvious to me that there would be, there was something missing at the product layer to really unlock the use of L L M's everywhere. And so that's why I've kind of motivated me to start building in the product space rather than doing research.

• 6:00Mario Gabriele?

  And and just to go back to that moment in the summer of twenty twenty two, Harrison, do you remember like what you guys were sort of talking about at that point? Like, what was the tenor of those conversations or sort of the the contours of what you were thinking and might be possible or might be interesting?

• 6:15Speaker 02

  So I think, and Stan, you should correct me on this, but I'm pretty sure I just cold emailed you or cold DM'd you on Twitter or something. So I can share a bit of my background as well because I think it leads up to this. But my background, so I studied stats and computer science, then worked at a fintech company doing more like time series stuff, but a little bit with some of the early BERT models on entity linking. And then I went to an ML ops startup where I was doing kind of like tooling for ML. And then I remember in like August, September, I was going to a bunch of meetups in SF. A lot of them were on generative UI, but they were more on like stable diffusion and image, things like that. But there were a few people doing cool things with language models. And I just remember being like, holy crap, these language models are fantastic. They're just

• 6:58Speaker 02

  so different than like the traditional kind of like ML models that I'd worked with before. And so then I think I was just paying attention online and Stan, I think, was tweeting a bunch about stuff that he was working on around early versions of Dust. And I think I just like either DM'd or emailed. And he was very gracious to kind of like respond to that and hop on a call. And I think, I think we had one or two calls and just kind of jammed on that and then kept on working. Yeah, that was the, I think that was the start.

• 7:29Speaker 00

  Yeah, exactly.

• 7:30Mario Gabriele?

  That's awesome. Yeah. So you talked about this sort of discrepancy between, you know, the power of the technology and the impact it was having in the world. And, you know, it sounds like that was sort of the, maybe the spark that led you to Dust. But like, more specifically, what was it that you were saying, here's the gap, or here's the real thing that needs to be solved? And like, here's how we, we go about doing it?

• 7:52Speaker 00

  To be very candid, at the beginning, it was not necessarily perfectly identified. It was really, this thing is great, nobody uses it. And you're like, what is happening? Where is the heat being dissipated? It must be dissipated somewhere. And it's unclear where. And I think it was at the product layer. And ChatiBID was a confirmation of that hypothesis. ChatiBID, in a sense, is mostly a product, as a bit of a model work compared to what was available off the shelf at the time, but not that much. It's just a very nice UI and the fact that it was free. Super interesting to think back about, and I don't want to like sidetrack you too much, but Character I was incredibly well positioned at the time. Character I, at the product, at the research team, at the model.

• 8:37Speaker 00

  Why is it ChatGPT that takes it all? Why Character AI didn't explode at the time? That's a very interesting and fascinating question that we'll be able to study when we start doing history of that period, I guess.

• 8:49Mario Gabriele?

  Oh, interesting. I appreciate the desire not to get too sidetracked, but I do think that's interesting. Like, what was, what's your sort of working hypothesis of why Character AI maybe wasn't the one to capitalize in that moment?

• 9:03Speaker 00

  So Character AI, I think, was speaking to a specific population. It was kind of weird in many ways. You would like start by create a fake character or you would create Elon Musk clones or clones of Albert Einstein or whatnot. And I think it created a complexity level even for the, like, the pure B2C audience. It was maybe a bit too high. It had a very strong community, which is the interesting bit. It was going very well. It's just I didn't just catch up the same way that ChatGPT did. Did you need the OpenAI brand gravitas for ChatGPT to emerge? Maybe, because AI was not completely unknown. I mean, most people wouldn't know OpenAI, but in the tech sphere and the journalistic sphere,

• 9:48Speaker 00

  I guess, people did know it. It's only hypothesis, I'm not sure. But Character was still a slightly, kind of, a bit complex, a bit gimmicky, a bit geeky product. And maybe they missed on that opportunity because of that, I know. What do you think, Harrison?

• 10:04Speaker 02

  I think there's some level of kind of like simplicity that just ChatGPT kind of like brought. It was just one chat box. You didn't have to, yeah, you didn't have to select who you were, who you were kind of talking to. Even now, like, I mean, this is, this is again, a bit of a segue. But, and I'm actually curious how you guys handled this at Dust, but in a lot of consumer products, you choose kind of like what model you want or things like that. And I think there is a simplicity in just like, you know, just one chat box type in. And, you know, now OpenAI has this, but at the time when they launched, I think it was probably just one model that they gave you kind of like access to. So you don't have to think about that. And then I had some friends at Character as well. And I remember when ChatGPT came out, they were very worried by that. And obviously, ChatGPT exploded more than Character, but it did kind of, it was a little bit of like rising tide lifts all boats. Like, there was just more people interested

• 10:49Speaker 02

  in kind of like chatbots and seeing what was going on. So I feel like to some extent, it's a consumer product. And so who can really kind of like, you know, know exactly why the culture's like geist catches on to one thing. But once it does, it's just, you know, explosion.

• 11:03Mario Gabriele?

  And so for you, Harrison, what were sort of the precipitating events to going from the ML Ops company you were at to really deciding, hey, there's something to be built here and the thing that I want to work on is LangChain?

• 11:15Speaker 02

  Part of it was my background's just in kind of like the building developer tools. Part of it is my background's in ML and I got really excited when I saw these models and was really like, oh, like, these are, these are kind of like amazing. So I was still at my, my previous company at the time and I knew I was going to leave. I didn't know what I was going to do, but I basically wanted to explore this area to kind of like figure out what, what I would do next. And my plan was to basically leave, take one to three months to just like figure out what to do next and then start working on it. But I was, I was sticking around, the CEO asked me to stick around for a few months to help with the transition. And so in that time period was going to meetups. That's when I reached out to Stan and wanted to just build some things, kind of like to get my hands kind of like dirty with the tech.

• 11:57Speaker 02

  And I remember chatting with Anton, who's one of the co-founders of Chroma, another kind of like developer tool in the space. He actually remembers this story better than I do, but apparently I had kind of like four ideas, which I was talking to him about. And, and one of them was LangChain. And then one of them actually would end up being LangSmith, which is our commercial product that we, that we build. The difference is LangChain, you can build kind of like nights and weekends. It's just an open source, kind of like Python package. And, and so I built that. LangSmith is more of an actual product. And so, you know, that, that came later, but basically just wanted to build stuff to kind of like get, get my hands dirty, released it, tweeted about it, kept on adding to stuff. A month later, ChatGPT comes out, keep on doing the same. By the time I end up leaving my company in, in early January, it's pretty clear that

• 12:41Speaker 02

  there's, you know, LLMs are going to transform how applications are built. And we, we think there's a lot of tooling that needs to be built around them to make it easy to kind of like plug them into applications. And so kind of like those two things, like LLMs are great, but then also we saw this with ML stuff and ML Ops, like there, there is tooling that needs to be built and a lot of that manifested in, in LangSmith and other tools that we're building.

• 13:04Mario Gabriele?

  It sounds like you had a lot of clarity around sort of like the early vision, both with Lang Chain and Lang Smith. Are there parts of that vision that have surprised you with how true they sort of have turned out to be, or parts that maybe you were making a bet on that you're like, oh, actually, the way the industry has developed is actually quite a bit different than I initially foresaw?

• 13:24Speaker 02

  I think like, high level, like, you know, maybe we had some clarity that we were right on, like, high level, like, we thought LLMs would be great and transform how applications are built, and high level, we thought there'd be tooling that would be needed to build around them, but like, that's very high level. I think a lot of the lower level stuff, like, we were figuring out as we went along. Maybe one thing to that effect, like, Lang Smith, which is our tooling for kind of like debugging and testing these types of applications, we see it being used by a whole set of kind of like builders, so not just engineers, and not just AI engineers, but product folks and subject matter experts and everyone coming together. And when we were building it, especially because of my background in ML ops, where

• 14:05Speaker 02

  in ML ops, the people using these tools were ML engineers, a small segment of the population, you know, I don't think we fully appreciated or anticipated how many different types of folks would be involved in the creation of these types of applications, and how therefore our tool would need to respond to these different folks. So I think like, you know, high level, I'd say we were, you know, I think we had some good insights, but like, low level, all the details, I mean, that was like, no way we could have seen all of that. That was just kind of like, you know, peeling back layers of the onion as we went on.

• 14:36Mario Gabriele?

  Stan, it sounded like you sort of maybe started with an acknowledgement that this was, you know, a very exciting space, but the exact sort of dimensions of Dust have been more emergent. Like, what does that journey look like for you? Like, what were the early bets you were making that, yeah, have been proven true and maybe have been a little different?

• 14:52Speaker 00

  Yeah, I think the early days of Dust is mostly me, as Harrison did, mostly me messing around with a product, and I think we really started building what Dust is today much later, actually. It's more like early 2023, when my co-founder Gabriel joins me. And there, at this moment, we decided to really focus on applying LLMs to Internal productivity and so not building. Dust was at the beginning something addressing the same kind of space as Langchain. Because the developers was the only person that you could talk to, basically in late 2022. But as ChattyBD comes out, it kind of educates the entire market. And now it opens up the opportunity to talk about L L M's and agents, not agents yet, but L L M's.

• 15:37Speaker 00

  And how it can impact work much more broadly. And so as Gabriel, my governor, joins, we really decide to focus on enterprise internal productivity. And applying that technology to how people work. This is a very long journey that we are only starting. I think the great thing is that there is so much to be done. And the form factor of what it's going to mean to work with agents is still pretty much unknown. And so it's been as well a very long iteration on many stuff. We started with the idea of with two main hypotheses at that time. So it's kind of the rebirth of Dust as we incorporated and start working together.

• 16:20Speaker 00

  And I think we start with two main product hypotheses was first, we need to have the company context. Because if we don't have the context, we won't be very useful for doing any actual work. There is a lot of work that can be done in wizard context, translating, creating ideas, doing research on the internet, etc. But for doing actual work at scale within the company, we want to have access to the company information. And the second thing was pretty simple is that, I mean, the company context is kind of big. So if you give it to an agent or to an L L M, it's going to be lost in it. Because the retrieval is not perfect, the models are not perfect. And so the second hypothesis was really the ability to create custom agents. So that you could pinpoint what context you really need for a specific task, which mechanically gives you much better results.

• 17:02Speaker 00

  And those two hypotheses were somewhat early in the markets, meaning that when we were pitching Dust in 2023. Everybody was kind of looking at us with weird highs. And we've seen the market open up to those ideas, which has been super exciting. I think the latest step function in there is probably the Toby Lutker memo. Which really made us to some extent go from the early minority to the to the to the early majority. I don't know, don't remember the the exact terms, but we we we went one step in the adoption curve. So we we kind of nicely positioned our place where the market is going now. But I think the the one thing is that it's a constant reinvention of what we are trying to achieve. Because it's for sure not the end state and we can dive into that.

• 17:44Speaker 00

  We have many, many opportunity to see that through different prisons.

• 17:49Mario Gabriele?

  I'm sure we will dive into that. And and just for listeners, you know, I'm sure most folks saw it. But the Toby memo was, you know, Toby from Shopify basically saying, if you're not using AI for your job at Shopify, you're in big trouble. And so you need to really be adopting it in a very aggressive way.

• 18:06Speaker 00

  Yeah, I think it was it was slightly more gentle. If I'm maybe I'm wrong, but I think it was don't come at me for head counts. If you haven't tried using AI first.

• 18:16Mario Gabriele?

  Yes, when I say in big trouble, I mean, less I'm going to come get you. But more like you're going to be left behind. And anyway, it was, you know, I think it was funny. We saw Toby write that memo. And then I feel like there was this mass memesis of like every CEO felt like they wanted to get their letter out into the public to say like, hey, we also really care about this and we're hardcore. It was very funny, but Toby Toby was early, I think. And yeah, I imagine that was a very validating moment. You know, we were sort of talking about this, the framing of Langchain and and us. But just to bring as many people along in this conversation as possible, like very crisply, what is the right way to think about what an AI agent like truly is and what it isn't?

• 18:59Mario Gabriele?

  Just so that folks, you know, have maybe a crisp wrapper for that that thinking, Harrison.

• 19:04Speaker 02

  An agent is an application where an LLM decides the control flow of of the application. And so I think that that's a bit vague. If you want to get more technical with it, I think for all intents and purposes, the way that a lot of developers think about an agent is running kind of like a while or for loop, calling an LLM to decide what to do and then and then taking those actions if they're actions. And Intel, it basically decides that it's finished. And I would say that, you know, like there, the LLM is very clearly deciding the control flow of the application. Like every single step is decided. I think you also have applications that are, you know, not as kind of like. Maybe agentic in some sense where the LLM decides maybe a few steps, but there's some steps that are hardcoded.

• 19:49Speaker 02

  Like maybe after A, you always do B. Or even when you talk about multi-agent applications, maybe you you you run one of these loops. And then after that, it immediately goes to this other application, which runs a check or or another agent even, and then and then goes back. So I feel like there's there's a spectrum. And Andrew Ng has a has a good way of talking about it that I like, which is like, rather than talking about whether something is an agent, let's talk about how agentic it is. And so that allows for this kind of like spectrum of of agenticness. And I would say, yeah, the the more an LLM is deciding what to do, the more agentic it is. And then probably at some point, there's some threshold where it crosses from, you know, a non-agent to an agent. And to be honest, I don't 100% know exactly where that is.

• 20:32Speaker 02

  But but I I like the idea of agenticness. But again, for for all intents and purposes, if you're talking to a developer, I would largely say they think of that as running an LLM in a for while loop and having it call tools until it decides that that's kind of done. Stan, is there anything you'd you'd want to add to that?

• 20:48Speaker 00

  No, yeah, I think the the the division is is is perfectly on point. I do like the the agenticness framing. That makes it enough sense. The big question we see in the market today, which is interesting, is agents versus AI workflows. It's a it's a different way of looking at the same question. There's a a lot of there's a lot of traction on companies that are doing AI workflows. We all know have heard about N80 and those those those kind of companies. And I think there's a there's a lot of value in AI workflows. And I used to use a definition of agents that I think apply to AI workflows, which is any program where one conditional is driven by an agent. But to some extent, that that agenticness framing is very interesting.

• 21:29Speaker 00

  Workflows versus agents is a massively interesting question, which is not answering your question and that we can revisit. But it's a it's a there's a lot of value in having workflows because it gives you much more control on what's happening. But to give you a sense, I think that long term, they they don't really make a ton of sense at the end of the day because we do. Fosia world where those agents will be actual coworkers. And I don't think you can really encode a coworker with a workflow. In the same way, you cannot encode Claude code with a workflow. We're really bullish on trying to help people create agents, not workflows.

• 22:10Speaker 00

  Very interestingly, which is the weird part, the agent is richer, but it's more risky in a sense. But it's also easier to build, which has a ton of value in the kind of a work setting. That means that anybody can build an agent, not anybody can build a workflow. A workflow is like typical make, an A10, Zapier and stuff. It's pretty easy. Most people will be able to interact with those products, but not everyone. And it requires a little bit of a kind of a learning and, and, and, and learning curve. Compared to that, building an agent is actually just describing in plain English what you want to do. And giving, clicking the capabilities of the agents should have. And it's, it, it, it's actually accessible by a much broader audience, which is, which makes it also very exciting.

• 22:53Speaker 00

  And so that's weird that the long term, most powerful thing is that's, is actually at the same time, the easiest to build to some extent. I'm, I'm obviously exaggerating because to build a great agent, you need evals and stuff like that. But to build a proto version of an agent and it's actually extremely easy.

• 23:09Mario Gabriele?

  Is it, is it a fair synthesis? And you guys can, can push back on this, but you know, if we're talking about sort of the workflow versus agent idea. A workflow might be more like, you know, writing out a recipe step by step. That you're sort of, you know, saying, here's what I want you to do. And an agent is more like creating a chef and saying like, please go cook me something. Is that, you know, roughly a way that someone might think about it?

• 23:32Speaker 00

  Yeah, I mean, it's a, it's a difference between McDonald's and a Michelin star restaurant, right? You know what you're going to get at McDonald's, it's very well streamlined. And when you're going to a Michelin star restaurants, you don't, you don't know what you're going to get. Because you'll be at the, the chef will be improvising with the situation and the food that has been available. It's, and it's again, a funny way to frame that.

• 23:54Mario Gabriele?

  Harrison, I saw your eyes sort of light up with a, with a question there.

• 23:59Speaker 02

  I mean, yeah, maybe pushing back on this a little bit. Like, I feel like to your point, Stan, I feel like you can often do the same things with workflows and agents. It's just the ease of how you describe it. Like, in an agent, it would all be in natural language, right? Like, you could have a recipe that you just put in natural language. Say, hey, do A, then do B, then do C. And, you know, it's not as deterministic, so it's not kind of like as safe, but it's way easier. And so I'm thinking out loud of this analogy for the first time, so I don't have a super strong opinions on it. But I do think, I don't know if it's so much different things, it's like different ways of accomplishing the same thing. And so maybe it's, I don't know, the difference of a person in McDonald's who's doing it

• 24:42Speaker 02

  all by hand versus using some of the pre-portioned things. I don't know exactly where I'm going with this, but I do think, like, I really agree with what Stan said, where like, there is a beautiful kind of like simplicity in how easy it is to build agents. It's usually natural language, and then you choose some tools. And I remember one of my favorite releases we ever did in LangChain, it must have been like the 13th release or something like that, where we took this idea of like the React agent, which is this great paper by Shenyu, and like was a little bit focused, and the examples they had in the paper were very focused on some like hot pot QA, some like Wikipedia question answering, like a narrow task. But I remember we took this and we made an abstraction around it where you just did exactly what Stan said.

• 25:25Speaker 02

  You gave it some tools, you gave it up a system prompt, and it would just do things. And I was like, holy crap, this is amazing. And I think like that is like that, there is this beautiful simplicity in agents. And so, yeah.

• 25:38Speaker 00

  And a very interesting data point here is that the React paper, which I invite everybody to read, and everybody will read that paper today, will look at it and say, what the fuck is that? It's so completely obvious. What, is it even a paper, say, say? It's funny, at the time, it was kind of a mind opening and people, we were so early in that technology that it was kind of a really interestingly mind opening paper, despite retrospectively it feeling extremely trivial and obvious with everything that's been built since then. And so that's really funny, a very funny exercise to all listeners to open the React paper and scheme through it quickly. That will show you, and it was a great paper, kind of a really engaging paper in late 2022,

• 26:23Speaker 00

  and that's how early we were. That's how, I said, so no clue we had about what we can do with elements.

• 26:28Speaker 02

  And maybe to that point, like when, when we launched it, and even right now in LinkedIn, I think we still call it like the React agent, but we're gonna stop doing that because it's just confusing to people. They're just like, what do you, like, what is React? Like, this is just so obvious, it's just running, it's like, you know, just taken for granted now.

• 26:43Mario Gabriele?

  That's amazing. You know, to, to maybe get a little bit more tactical with it, like, what are the, the different use cases you guys see for agents at the moment that are sort of most productive? And, you know, Dust is obviously creating a lot of those sort of for companies, you know, deploying them. But yeah, curious, like even within that, where are you seeing the most leverage for, for businesses?

• 27:06Speaker 00

  The list is long. We, we're taking the, we have a pretty horizontal, I mean, we have a completely horizontal product because we believe that it can be applied in so many places and that there is value of having one platform for everybody to share, creating and sharing agents and having agents interacting with others in a business setup. I think that I can only give you examples. It goes from extremely simple Slack thread to issue creation, and that's not completely trivial because at Dust we have something like three or four different types of issues, which goes in different types of projects, given the difference shape or, or type of discussion that needs to happen. And having an agent that helps you go from a thread, a thread Slack where there's a

• 27:48Speaker 00

  discussion, where maybe there's a bug that is emerged, there's a place of use where to put it when, when that's a bug, but when it's a decision that needs to be taken and it's to be moved into a different place and, and taking a few actions around that. And so streamlining that process of having something happening organically on Slack as an example, to following a workflow that makes it represented in the canonical way that a company represents that artifact that is being discussed on Slack is a very general use case that I covered for issues on GitHub, but that it can be done for many other stuff. Whenever there's a sales transcript, there's an army of agents that gets kicked in for providing feedback for autofire,

• 28:32Speaker 00

  filling in Salesforce, for extracting product interest and going to create comments on Notion pages related to the product that has been discussed during that transcript. And so there is so much that would require, so much stuff that would require human work that is now being doable by agents that is really interesting. The most interesting places is for the things that no human would ever do because it's just too much work. So for any sales transcript, being able to put a comment in the right product document on Notion about the fact that it's been discussed, the link to the transcript and the kind of a one sentence summary of what has been said is something that never happened before because nobody was there to make it happen.

• 29:17Speaker 00

  Nobody had time to review all the transcripts. And so those new use cases are almost the most, the most exciting as well for me.

• 29:23Mario Gabriele?

  Harrison, I'm curious, you know, maybe how you guys use agents at LangChain internally? And maybe some of the use cases you've seen that you've found particularly powerful?

• 29:32Speaker 02

  Yeah, so internally, we use agents in a few ways. So we, so and they kind of map with like the, the big use cases that we see out there as well. So like, we see customer support being a big use case. We've built slash are building an internal kind of like customer support agent to help us with a lot of those inquiries and responses. Coding is a massive use case. We both use and build some kind of like internal coding or coding adjacent agents for kind of like responding to issues, things like that, managing discussions. I personally use an agent to kind of like monitor my email and draft responses and flag things. And, and, and so that's, that's probably the one that I use the most.

• 30:17Speaker 02

  I mean, we use off the shelf and internal kind of like versions of deep research agents. Like that's been a huge kind of like style of things that we've seen pop up. Oh, we use some for marketing as well, like marketing's a fantastic use case. And so we use some for translating some of the blogs or things we do into tweets or LinkedIn posts. And so I think, I think those are a lot of the big use cases that, that we use internally. And I think they generally match up with what we see people building in, in the industry.

• 30:43Mario Gabriele?

  You've written, I think it was a blog post where you've talked about how we interact with agents and how that might change from sort of typing, prompting these agents with voice or text and shifting towards more of what you call like ambient agents that require less of that. Why do you see things going in that direction? And maybe you can share a little bit more about what an ambient agent really looks like.

• 31:07Speaker 02

  Yeah, absolutely. And I'm super curious to hear what Sam has to say on this as well because I think he mentioned something about thinking about how we interact with agents as part of the core mission of Dustin. That's one of the things that I'm a little bit jealous that we don't get to do a lot of that link chain because we are more developer facing. So we do think about it and it does absolutely inform like what tools we build, but not nearly as much as folks building products in the space do. But I mean, so far, the dominant UX for agents has kind of like been chat. And I think if you think about it from first principles, that actually does make a lot of sense. Like it puts the human in control. It's very human in the loop. You can, not only does the human initiate it, but the human can see what's going on because you can stream back results.

• 31:50Speaker 02

  If the agent wants to do an action, you can have the human kind of like immediately approve it there. So you can kind of like have some sort of like approval for dangerous actions. And it's relatively fast for the most part. But I think like, you know, people have been saying that chat won't be the only UX or the forever UX. And while it has actually lasted a lot longer than maybe people saying that would have initially thought, I do, I do think it's interesting to think of what besides chat is out there and also like what some of the downsides of chat are. And I think some of the downsides are also some of the things that make it good in some cases, namely like you have to kick off all conversations. So if you want to run it over kind of like 1,000 or 10,000 things, like that's, that's

• 32:35Speaker 02

  a little bit tedious to kind of do. And also because you generally expect to be in the moment, they can't really take that long. Otherwise you get a little bit bored and maybe switch. And I actually want to come back to that point because I think with like deep research and some of the coding agents, we're starting to see some like that's starting to happen. But it's still, anyways, I'll come back to that. And then the other thing is like, yeah, I mean, Just based on inside an enterprise or inside a company, you have all these events happening. And rather than like copy paste an email and then take that email and go and you know, put it in chat, wouldn't it be nice if that would just kick something off automatically? And so I think, I think the email assistant that I use is actually a great example of this. It just monitors my email inbox. It just gets triggered by these events and then it goes and does something. And if it wants to take an action that I, that I deem kind of like dangerous enough where I want

• 33:19Speaker 02

  to approve it, which right now is, is scheduling a calendar invite or, or responding to the email, then it presents it to me in some way. And I think they're like, what is that UX? What does that look like? Is that just a draft in my email inbox? Is I, we have a concept of like an agent inbox, which is a dedicated view for this. Back to the original question, like ambient agents, we define as agents that, that listen to a stream of events and then act on one or multiple at the same time. And I think crucially, these are not necessarily autonomous agents. They're still kind of like have some human in the loop at some component because I think that's still necessary for, for enterprise adoption. I have some other thoughts on, on kind of like the deep research and coding style agents, but I'd actually be curious to hear Stan's thoughts first because you work on actually delivering these to a lot of end users.

• 34:00Speaker 00

  Totally. I think basically the, the way we, we see it indeed is that the conversation interface has been the, the interface. We, we hypothesize, and maybe it's never been the case because indeed it's survived for a long time, that there's going to be a fork probably in the typical UX UI that means working with agents. As deep research agents take longer and longer, as agents are being triggered with you may not have the loop, you probably want something that looks like more like a comment center than a list of conversation. The conversation paradigm will probably make sense on the, in the B2C setup for a much longer time because in the B2C setup, the, the truth is that your agent is your executive assistant.

• 34:42Speaker 00

  And so you have one stream of conversation with them or a couple of different streams. But when you think about the enterprise, there's going to be agents that take a lot of time, agents, conversation with agents that involve multiple people.

• 34:55Speaker 00

  That's something that the market hasn't even started exploring much, right? I mean, we don't explore it much yet because I think we consider them as toy. And but the more powerful the use case will be, the more meaningful it will be for people to actually interact. Multiple people interacting with the same with an agent or with multiple agents. Completely aligned with your vision, Harrison, of the ambient agents. I think the first step, as you described it, is really to have more of an inbox paradigm. When you interact with those agents, having agents that are being triggered by external events. And I think the crazy idea is agents that are not necessarily mechanically triggered

• 35:38Speaker 00

  in the sense that when these do that, when these actually execute, but having agents that are just skimming through what's happening inside of the company and maybe ping you with offers to give you, to provide you some value, which is obviously what the end state should be. When you think about all multi-agent systems, I mean, today, we obviously don't see that in the enterprise. But you could imagine giving a very high level project to a set of agents and just let them walk and organize for delivering the project on their own and give you a report multiple days after. I think all of that is completely uncharted territory. And so, but yet, it's obviously the end state. And so that's why it's so important and so exciting to be working in that in that space.

• 36:23Speaker 02

  You mentioned a command center, a stand. Like, is that in the product now or is that a future kind of like direction? Because I love that idea, but I would love to see like, yeah, I would love to. I would love to see what that, I don't know what that looks like. I would love to see what that looks like.

• 36:35Speaker 00

  No, I don't know what that looks like either. I think the first step will obviously be like what you've been building internally. And I think you've shared on some of your blog posts, but then what you just mentioned, a form of inbox is just, is already a first step in that direction, obviously. The weird thing about the agents is that this, they also, I mean, the APIs are so biased toward a conversation.

• 36:57Speaker 00

  That often you're like, do something for me and you have that agentic loop that can last for a very long time. But at the end of the day, you still have an agent message, which is kind of weird because that means that you somewhat have a cap on the interactions through the agentic loop. You can make them very long, but you're also, you're going to be exhausting your context at some point, but that's not necessarily completely an issue. But at the end, the whole system is kind of, is kind of a post train to give you an answer. Yet, when you think about agents working, you just want to say, go do the work for a day and ask me questions if you have any, but don't give me an answer in thirty minutes. I just want something delivered in one day and ask me questions if you have anything.

• 37:40Speaker 00

  But yet, there's no good system for interacting with the current shape of those agents for doing that. You could think about multi-agent stuff, et cetera, but the ecosystem is not there yet, I guess. And so we still, even at the API and post-training level, a little bit bound to be staying close to the conversation all in the face. But I'm sure that we'll see kind of stuff emerge around that.

• 38:02Speaker 02

  When we were building at the start, like messages weren't a thing. It was just text in, text out, and then like OpenAI, I think it was, it might have been three point five or maybe four that they released and it was only the chat message API. And I remember talking with someone from OpenAI and it's like, so are you going to release like the non-chat message thing? And they're like, ah, we don't know. And they ended up not. And now everything is just like that. And that kind of happened. Two, what's also really annoying about this is there isn't really like a official schema for what messages is. Like OpenAI has their kind of like input output schema, but that's different from Anthropic's, which is different from Google's. And like, you'd think if this was like, this is like, you know, the base thing, which has how we interact with these. I, you know, I wish it was a little bit more standardized what that schema is, although it's constantly evolving as well.

• 38:47Speaker 02

  So tough to, always tough to do that. And then the third point is maybe like, so all the chat agents to date have been kind of just like synchronous agents. And that like you just chat in the moment. Now with deep research, you have things that start sync, you start with a chat, but then you go to this deep researcher. Then maybe you actually come back to synchronous at the end. And I think for some of these ambient agents. You could almost view them as async running in the background. But then at some point, like you said, they ping you with something and then they become synchronous. And so I think chat is a pretty good like form of synchronous communication. And then like, you know, what what async means is like, maybe that's hidden a little bit through some like context or prompt engineering. And it's just like, by the time it surfaces to the user, it's just all a message. Because that's the dominant form for synchronous communication, at least.

• 39:31Mario Gabriele?

  This episode is brought to you by Brex. Fred Adler, the influential venture capitalist of the 1970s, was known for displaying decorative pillows in his office that featured a signature business philosophy. Corporate happiness is positive cash flow. In today's post-Serp environment, Adler's wisdom feels particularly relevant as founders need to make every dollar work harder. That's exactly what Brex delivers. Their modern finance platform was built specifically for startups like yours and designed to help extend your runway when capital efficiency matters most. With Brex, you get global corporate cards with up to 20x higher credit limits and no personal guarantee required.

• 40:14Mario Gabriele?

  Their banking solution has no minimums and no transaction fees, while letting you earn high yield from day one with same-day liquidity. Best of all, Brex knows you were born to build, not juggle spreadsheets and finance tools. Their AI-powered platform brings cards, banking, expense management, and travel all in one place. It's simple, scalable, and designed to get you back to what you do best: building. More than 30,000 companies, including one in three US venture-backed startups, trust Brex to help make every dollar count toward their mission. Join them at brex.com/mario.

• 40:54Mario Gabriele?

  Also, it sounds like there's just this layer of proactivity that, you know, you're suggesting might be different, that you're sort of saying, here are the the goals that we have as a business. And actually, as long as you're sort of have the sufficient context and, you know, enough power, you can start to say, hey, by the way, you should consider doing something as little as writing this tweet to boost the numbers that you want to do. Or to something as big as you should consider, this new product that might be really important over the next few years. What are the sort of major limitations to a more ambient model today? Like, how far are we from a true command center world, where maybe you're really ushering out a swarm of agents per person

• 41:38Mario Gabriele?

  and having them sort of monitor and think for you and do this deep asynchronous work on a regular basis?

• 41:45Speaker 00

  I think reliability is obviously a limitation. It's still mind blowing to me how dumb those agents can be in pretty obvious situations. And yet, F* get a Naimo gold medal at the same time. It tells a long story about the importance of data, the importance of pre-training, the importance of post-training, and how there's been focus on code, on math a lot. And yet, on different places, there is some, obviously, some gains as well. But so many cases where they're like, damn, you're being so silly there. You can solve very complex math problem, and yet you don't understand from the context that it is two men speaking together or whatnot. Anyway, so I think it is the main blocker.

• 42:30Speaker 00

  And so, as an example, I wanted to share that. I think a way to, a different way to think about working with agents is that could be a transition towards the very long ambient agents is the concept that I really like, which is the concept of a work plan. And if you ask me, I don't understand why linear and all the kind of ASAN stuff isn't doing that aggressively. But you can imagine a work plan. You have a very high level task, and you start splitting it in smaller tasks, and you start splitting the stack in smaller tasks. And once you start doing that work, you can do it assisted by an agent. Okay, I can do it yourself. And then you can start delegating or discharging those tasks to agents. And those agents start working on the task, come back to you, and like, no, not quite yet. And eventually, you start clicking the tasks that being done. Maybe it's you, maybe it's another human, maybe it's an agent, maybe it's a bunch of agents.

• 43:15Speaker 00

  And so there you kind of have a nice mesh between the conversation and the kind of ambient agents. It's not at all ambience to begin with, because you kind of develop the work plan and discharge to agents as you as you go. But the better the agents are, the more they'll be taking of that work plan task. All the way to eventually, maybe someday, defining the work plan, speaking the task, discharging to other agents, etc. And so I think the there's this even in in the current world where we have very deep limitation in the reliability of agents on some task, which make the presence of a human to monitor what's what's going on kind of very important.

• 43:53Speaker 00

  I think there is many products that face we can imagine that will start to, you know, mesh between the sync interactions all the way to more I think through the ability to probably introspect what has been happening.

• 44:08Mario Gabriele?

  Does that chat with you? What you think Harrison is the major blockers?

• 44:11Speaker 02

  Yeah, I I would agree with that. I mean, I think like reliability of individual agents. I then think there's a lot of work to be done at kind of like the UX layer. And then I'd also say like, and this kind of gets to the reliability aspect, but just like learning slash memory is also interesting as well. Like, there needs to be some like that. That's what we as humans do. And so that's maybe a little bit further out. But I do think that's a component. I also think like, I think code often leads the space just because the models are really good at it. And so I think if you look at like Claude code, like that's a great example where the model got good enough. Okay, so reliability is a little bit better. They did a good job of writing up a CLI and giving it access to some tools, so some great context engineering there. And then you start to see like some interesting UX things happen.

• 44:55Speaker 02

  So I think I think there's an open source project called like Taskmaster or something that like, you know, keeps an eye on like five or six Claude code things that that happen kind of like in the background. I think they released a view to kind of like see kind of like usage of Claude code. And Chip Hewen released something as a way to debug like the errors that Claude code kind of like made. So like now that we get these like more like, I think that's the first example, or that's a, yeah, one of the first examples of these really like long running, kind of like more autonomous agents. And now you're starting to see a bunch of like interesting. Kind of like command sentry type vibe things coming out for how to interact with them. But it's still, it's still really early on, but I, I, I like to look to, to code for an example of like where, where the space is generally headed.

• 45:39Speaker 02

  Just because I think it's ahead of, of the other verticals.

• 45:41Mario Gabriele?

  Given how Dust operates, I imagine you have a, a specific opinion on this. But when you think about how agents play out over the next few years, do you think that like it's unequivocal? That there's going to be really many, many specialized agents or over time? Do we just sort of start to converge into, you know, a super agent that has enough context on work and life or whatever it is? Yeah. Are we, are we heading towards a, a true multitude or, you know, an oligarchy or, or, or one true, true ruler?

• 46:13Speaker 00

  That's the big question. We don't have a clear answer on that and we're trying to stay very humble with respect to that question. We started with many custom agents and it was a clear, good decision at the time, given the state of the models. As the models are getting better, there is an undeniable force towards higher level agents. Until the agent is, is, is, doesn't have a really functional memory so that it can interact with humans, learn from them, being coached and understanding how the company operates. I think you're still going to have the need for custom agents because if the agent doesn't have a good episodic memory, in a sense,

• 46:55Speaker 00

  it's going to be very hard for them to learn that this data is routed and this data is fresh. I mean, in every company, you have data that is not up to date and data, data that is good and data that is bad. And you have ways of doing stuff, et cetera. And so today, having custom agents lets you point to the right data, explain the right process so that you don't have to do it each time. I think the state of memory of agents doesn't scale us to a point where you could have just one agent and it's going to learn it all. And also. Kind of feels weird that you would have to teach your agents. And there's also weird stuff in a, when a company gets somewhat big,

• 47:39Speaker 00

  you even have contradictory ways of doing stuff within the company. Team A will do stuff this way and team B will do stuff this way. And so now it begs the question, where's the memory? Because teams will be competing for the same memory slots in the sense of doing the thing the right way. So there's still many questions, even if you assume a really great, perfect agents, there's still a ton of, a ton of questions. To be answer your question, I don't know if it's the end states. I think the level of abstraction of the agents in general will increase. And so the number of agents being necessary to walk and to do work will probably decrease. There will be probably a convergence toward one, but it's very unclear when that's going to happen.

• 48:22Speaker 00

  And I think we're trying to keep our, our finger on, on that, that trend.

• 48:26Mario Gabriele?

  So you do think eventually there'll be a convergence towards one workplace agent?

• 48:30Speaker 00

  No, I'm sorry. I'm saying that there's going to be a, an increase in the abstraction level of agents. And so a decrease in the number, I don't know if it's going to converge towards one. I guess, I guess we'll see it's converging, but maybe there's going to be a ceiling.

• 48:43Mario Gabriele?

  10 versus 100 or whatever it might be.

• 48:45Speaker 00

  Exactly, yeah.

• 48:47Mario Gabriele?

  Does, do you take the same position Harrison or do you see things a little differently?

• 48:51Speaker 02

  No, I think I largely agree, maybe like a few kind of like, you know, thoughts as well. Like one, like generally, like what, what does it even mean to have like, like what are, like, what does it mean to have multiple agents? Like, how are they different? And generally it's, it's the prompts, it's maybe the model, but mostly like the prompt and the tools it kind of like has access to. And so sure, in the, in the limit, you could maybe have like, you know, one, one agent with every single instruction for how to do everything at the company and the system prompt and all the tools there under the sun. That's definitely not what we see right now. Maybe it will go towards that direction or towards a smaller number of agents. I think what we see more now and is maybe an alternate view is like,

• 49:33Speaker 02

  there will be one agent that a user at a company interacts with, but under the hood, there will be. Be many subagents that it can call out to or route to or use to and and those have like the specific instruction that you know, like when we talk about people for building is like write down a standard operating procedure and figure out what tools it needs and then that's your agent. So maybe there'll be like, you know, one kind of like central supervisor agent that can interact with all these other agents either by and and now we get start to get into multi-agent stuff and that's very, very kind of like early on, I would say, but there are some initial ideas of how to do that. I mean, even if you look at some of the coding agents, going back to kind of like looking at code, uh, Google Jewels is kind of interesting. It has kind of like this chat based synchronous agent that kicks off other async kind of like background agents. And I'm assuming there's some difference in the system prompts and tools that has access to. Something

• 50:18Speaker 02

  like that, I think is very, very reasonable. And most people, even right now, are kind of building towards because people don't want to have to choose like, oh, yeah, I have hundreds of agents. Like, no, they just want a chatbot. It's this, it's a simple kind of like, you know, approach to that. But under the, under the hood, at least right now, and and even for the foreseeable future, I think they'll still be like relatively specialized.

• 50:37Mario Gabriele?

  It's like having, you know, your agent that is your VP of marketing, who's also managing the agent for your social media marketing, your performance marketing, your brand marketing, and you don't have to worry about, hey, I'm trying to figure out, you know, how to do my performance marketing. Here's the exact one I have to go to sort of thing.

• 50:52Speaker 02

  I think that's exactly right. I think like, um, one thing that we do, and I genuinely don't know if this is good or bad, but I feel like we often like anthropomorphize how we interact with these agents. And on one hand, it might be good because like, yeah, that's how we are used to communicating and that maps to our mental model. And that's a good, like, all these like context engineering, which is like, you know, the, the topic of the month is just communication, right? But on the other hand, like, these things are different than us. So like, why should the way that we communicate be the way to keep? So like, I genuinely don't know if it's good or bad, but like the analogy you just made, like, I think that's what we often do and other builders often do to try to figure out what the best way to organize and communicate across these agents are. Again, for better or worse.

• 51:35Mario Gabriele?

  Here's a question I have that maybe goes beyond the paradigm of the agent, which is, how can we make sure the agents and AI in general is doing properly useful work when we see so much of this sycophantic posture from a lot of their responses? Like, you know, something I worry about when we see a company full of agents is like, will you really have a VP marketing, VPN, whoever, who's really able to think critically about this when there is just this, you know, reflexiveness that is so pleasing? Like, do you see a solution to that question anytime soon?

• 52:12Speaker 00

  I don't have a solution to offer like that, but I feel, I mean, my, one of a small side research project I'd love to work on, I don't have the time, but if I had time, I would play with that, is probably to try to have agents debating against each other towards a goal.

• 52:32Mario Gabriele?

  Like adversarial?

• 52:33Speaker 00

  Not necessarily adversarial, maybe more like a research community. They share results and then you have something like a hacker news, a hacker news system where the things that are the most cited go up and it's a clear objective of the agents to get ranked high and they try to push towards getting some answers with that by trying to follow some form of notion of truth, which is obviously a whole, I mean, you have no guarantee that they would do it or whatnot. But I think there's in the multi-agent setup, there's probably a new dimension that it creates that can probably alleviate that problem because you'll have an opportunity to prompt agents to be actually a little bit adversarial to other agents, providing a very, as you mentioned, very reflexive response that try to please the user.

• 53:18Speaker 00

  And so I think there's a lot of stuff to be explored there. Obviously, we are light years away from productizing this kind of stuff. We're light years away from productizing this kind of stuff because the state of the market is light years away from even those kind of questions, which is interesting. But I think there's a lot of stuff to explore in that direction.

• 53:40Mario Gabriele?

  Harrison, anything that comes to mind for you there?

• 53:41Speaker 02

  Yeah, prompting these agents to have different points of view is practically speaking, what I think is feasible now. And then I also imagine some of kind of like these, you know, issues will get handled through better models that just come out from the foundation model labs.

• 53:55Mario Gabriele?

  Are you seeing folks do some of that prompting to do some of those sort of like prompting different views and sort of, you know, strapping that together to have like the hacker news style ranking or whatever it might be?

• 54:06Speaker 00

  I mean, I know there's probably a ton of teams over the world working on those kind of ideas. OpenAI obviously has a multi-agent team. It seems like the IMO result comes from the multi-agent teams. They say they have a special model. If you ask me, I would say that it's probably a multi-agent setup, but they do exactly those kind of sheets. And so I think many people are exploring for sure, and that makes sense. But it's still definitely in the realm of research at this stage, I would say.

• 54:31Speaker 02

  I think we see like very like, um, simple and naive versions of that, or a simple version of this is just like reflection or critique on an initial thing. And so like, I think like a, a, a, a pattern that we sometimes see is like, yeah, have one agent or one LLM generate something and then, and then give some feedback on that through whatever. I mean, this is, this kind of gets into like some of the reward systems that actually go into RL. And so like for code, it's kind of easy. You can run the code. So you actually don't even need another agent to provide this kind of like other point of view, right? You could almost view it as like, hey, this is the system's point of view where your code doesn't compile. That's just like a fact, right? But I think you can imagine doing stuff like this for, you know, essay writing or something like that, where you have, you

• 55:13Speaker 02

  have kind of like one agent that, you know, reviews it or, or gives some feedback. I think right now it's a little bit more researchy, unless you have kind of like these verifiable kind of like rewards almost that you can feed back into the agent as it's running. Like evals in the loop is kind of like what, what we call them. And so you can add these checks from, yeah, like running codes, kind of like the most obvious example. We're working on an internal, uh, coding agent. And as part of that, we're experimenting with having kind of like a separate agent that kind of like decides whether it's, you know, done with a loop. And that's a little bit different. It's not like as adversarial, but it's kind of just like delegations of concerns almost, or separations of concerns. Um, and so I, I

• 55:57Speaker 02

  think you could, I think you can view some of the, some of the, some of the stuff that people do in this vein, but it's very kind of like brute force in some way or like simplistic.

• 56:07Mario Gabriele?

  I'd love to zoom out for a moment and talk also just about what it means and what it's like to be building in AI at the moment and some of the specific dynamics that founders have to face. I think one of them is really just how fast the fast following is happening. You see any good idea, there will be three, four, five, however many folks chasing that very, very, very quickly and able to raise considerable amounts of money. As you've gone about building your businesses, like, how do you think about protecting against that, building in defensibility where you think you have a real chance to build a moat? Yeah, Stan, how have you thought through that at Dust?

• 56:51Speaker 00

  I mean, so building in AI is a clusterfuck, that's for sure.

• 56:57Speaker 00

  So basically, for the past many decades, the technological substrate has been extremely stable. For the SaaS, let's say, for the SaaS decades that were behind us, it was JavaScript and Postgres. I'm exaggerating a bit again, but extremely stable technological substrate. So when you were building something, you knew that the foundations were not moving, so you could describe where you were going. You could build an image of your vision. So you had the vision and you had what you wanted to build to realize that vision. And today, one very specific thing that you faced as a founder building in AI is that you have what I call the fog of AI. And the fog of AI is the fact that the foundations are moving

• 57:40Speaker 00

  very quickly. And so you have to have a vision of where you're going, but you cannot paint it, because if you paint beyond six months, whatever you're painting will probably be shattered by the foundation shifting towards a different, I mean, the spacetime of the ecosystem shaping itself in a different ways. And whatever you were painting will probably be not true. So that means that you have that fog of AI at six months, which is very, interestingly, very problematic for like, building a high-efficiency organization, I find, because alignment is one of the things that makes organization extremely efficient. And here, you don't have that kind of continuity between the current products and the vision.

• 58:24Speaker 00

  You cannot paint that continuously, because you have that fog barrier at six months, which means that you have to jump from the roadmap for the next six months to the vision. And that makes... Alignment of the team, a challenge that is interesting in the way people think about where the product will be, how they prioritize stuff. You want all of that to be as autonomous as possible. And that makes it a really strong difficulty compared to what I've seen. I've been lucky to be at Stripe and at Stripe, we had a very clear alignment because it was a simple developer-centric product through an API. And so you had that very strong alignment internally that allowed for the organization to grow and be efficient without a lot of processes.

• 59:03Mario Gabriele?

  This is the lesser discussed AI alignment problem.

• 59:06Speaker 00

  Yeah, I find that, I find that, I find that one of the most challenging part of building in AI.

• 59:10Mario Gabriele?

  I bet, yeah. And so how have you thought about building the defensibility piece for Dust? Like where?

• 59:16Speaker 00

  Oh, yeah. So, sorry. And I mean, we've, we've, we've managed to build a product that was slightly in advance of phase compared to the market. And now we see the market move into that. And as the market moves to that, the, every big players in the market is waking up to it. So you have Salesforce with agent force, you have agent space of Google, I mean, everybody is working up to it, to it. We've had two years of building a product as best as we could that is really creating us a defensibility today, meaning that our product is probably in a better state than most of those big players are shipping today. But they're also big players with many developers, so they eventually ship the thing. I mean, we can trust that.

• 59:56Speaker 00

  And so I think it's always a question of trying to build a few, some time in advance.

• 1:00:02Mario Gabriele?

  Yeah.

• 1:00:03Speaker 00

  Which is completely contradictory with what I just said before. But that's part of the customer of building in AI is that you must be building two years in advance, even if you don't really see it yet. And so that's a real challenge. Being in the first is more a common center. We don't know what it looks like, being in something like work plans and stuff like that that we discussed on, on the podcast. I think we don't know exactly what it looks like, but you have to be investing a lot of resources there because you, you have to build conviction that it's where it's going to be in the future and you want to be building it now. And as you do those phases, the, the bigger you get and the more gravitas you get and the more, the more resources you get. And you have maybe a chance of arriving to the OpenAI and the Microsoft and the, and the Google of the world. That's for us.

• 1:00:48Speaker 00

  Or with you, Harrison.

• 1:00:50Speaker 02

  Yeah, I mean, I think like I agree with everything you said around just it being a chaotic time to build. I think like, you know, execution is a moat and execution speed and like that's, yeah, to the point of, you know, building fast and thinking in the future. Like, yeah, I think like we, you know, I think honestly, the team that we have at Lingchain is fantastic, and that's that is a big moat we have. And I think we do execute like really fast and really efficiently. I think like, you know, a little bit, maybe more kind of like in the details on that or other than that, the fact that there's so much going on in AI can actually be a blessing in some sense as well, because competitors will get distracted by other things as well. So they might see something that you do and be like, oh, that's cool. But then they see something else that someone else does. And they're like, oh, that's cool as well. And so

• 1:01:34Speaker 02

  they'll, so I think like, really trying to like, understand the problem and having conviction in that and like, building towards that in like a, I actually think like, just like, understanding in general is actually very hard. And so from like a product point of view, just like, understanding what you're building towards and having a consistent kind of like product strategy in that, or product experience in that. And then, you know, the features that you add, someone else may be able to copy them. But if they don't have that kind of like holistic understanding, they're not going to do as good of job. And it's going to show up kind of like at the margins. And then the other thing that I'd say is like, I think a lot of the early things we did were around that kind of like, understanding the user experience and building towards that. And now we're starting to build, we're starting to try to figure out like, okay,

• 1:02:19Speaker 02

  what are the kind of like, deeper technical bets that we can make that like, these other things like all kind of like boil down to in some way. And again, like, and there's like, two or three in particular that like, we're kind of like thinking about. And that's not that many, right? But we will way many more things, but they all kind of like come back to this. And so if you're looking from the outside in as a competitor, you might say like, oh, my God, they do like 100 things. But there's really like two or three kind of like tech, deep technical things that we're betting on. And that's just, you know, having conviction and kind of like, I think it's tough because you do need to be moving fast, but you also need to kind of have some sort of consistent kind of like conviction or consistent kind of like, technical bets that you're making to build up that that that moat that can't just be replicated

• 1:03:03Speaker 02

  in like a week. So it's a, yeah, it's a crazy time. But that's how that's how we think about it.

• 1:03:10Mario Gabriele?

  You know, at the time that we're talking, not long ago, we had sort of the windsurf drama opera of You know, OpenAI buying them, then that falling through, then, you know, management getting picked up by Google and then, you know, cognition sort of taking the the rest of the company and sort of saving the employees from from being left without anything from what was looking like a really massive acquisition. Do you think that's like a version of M and A that we're going to see more often? Was there enough of an organ rejection from the startup community towards that practice? That like, hopefully, we don't see that as much? Or you're seeing, you know, talent sort of respond to these kinds of behaviors?

• 1:03:54Mario Gabriele?

  I'm, yeah, just curious for your take as founders on the ground and and how, yeah, those those sorts of things are are changing the field, perhaps.

• 1:04:03Speaker 00

  Weird, it's probably going to be controversial, but who cares? I much, I think I'd much rather prefer the the whole wins of setup than the scale AI setup. Why? The wins of setup is actually an acquirer and there's been a career forever. And acquire have always been selective on the people they take in. The the weird thing about that acquire is that the the amount is just doesn't make sense. It's just way too big. And so the amount makes it like makes it really not great. Because you really paid 2.5 billion for some folks. Why don't you get all the folks, right? Generally, the acquire is when the company is dying in a sense.

• 1:04:45Speaker 00

  And it's an event that is great because it lets you join a bigger team, but it's not with massive amount of money. So the kind of stardom system of AI, etc., makes those accquirers completely weird. But I'd much rather have that than scale AI is. It feels to me slightly more weirder because it's both the CEO acquire here, but at the same time, a not fully completed acquisition, just a majority stake buy, which I don't know where what were the dynamics for in terms of returning to the employees. And it feels kind of a bit, almost even a bit more complex and I as a sense a bit more perverse.

• 1:05:27Speaker 00

  But obviously, I all that being said, I do think that it was really not acceptable for the employees that were seeing a bunch of folks leave for 2.5. I don't remember numbers, but for billions of dollars and be left with something like, what what are we doing, guys? And then somebody tap you in the back and say, you've got a running company. That's great. Go get it. That made the whole setup weird, but at the same time, if you look back and forget about the amounts, it was still mostly in acquire. It's just it was weirded up by the massive amounts involved.

• 1:05:59Speaker 02

  This isn't the first time it's happened. I mean, like, character inflection, adept all had versions of this as well. I also think like you're seeing in the markets, like, there's just this insane price for talent that's going on with, you know, meta and, you know, the rumored salaries and stuff that they're paying people to try to get from open air. And so I think it's like a really just like crazy time in the talent market. And I think that's manifested in a few ways, including these offers, but also including these kind of like aqua hire acquisitions, whatever, or faux aqua hires, whatever you want to call them. I feel like the windsurf news is new enough where I actually haven't had that many kind of like detailed conversations with folks about it on the ground. I think it happened, what, last week or something like that, or a week and a half ago.

• 1:06:42Speaker 02

  Yeah, I mean, I don't really know how it will affect things going forward. I do think that it's, you know, I imagine that's not what the founders had in mind when they started the company or even, you know, six months ago or something like that. And so I don't think it's a great situation at all for kind of like the employees that were left there. I also don't, I don't know, like, I spend a lot of time thinking about why I wanted to start a company. Before I started a company, I came to the conclusion that I wanted to build something great with people I enjoy working with. And I feel like, you know, I feel like not only do we have that here at Langchain, the chance to build something great, but also like I really enjoy everyone that I work with here.

• 1:07:24Speaker 02

  And so I think like that's personally kind of like what motivates me. And so I think I have a tough time kind of like seeing kind of like myself or Langchain going that route. But, you know, I also think if you asked the Windsurf founders a year ago, they probably would have said the same thing. So I don't want to, I'm not, I'm not here to judge anyone, so.

• 1:07:42Mario Gabriele?

  Yeah, I think your point about the, just the intensity for talent right now is like, you know, really such an important one. As startups, like, how do you think about where your competitive advantage is in such a hot talent marketplace? Like, where have you found you're able to compete most effectively? And get the folks that are like perfect for your particular mission?

• 1:08:07Speaker 00

  And for us, we're building from Paris, so we have an easy

• 1:08:12Speaker 00

  easy on this one.

• 1:08:15Speaker 00

  I think there's still some competition and we, but it's nowhere close to SF and I think we are capable of

• 1:08:25Speaker 00

  creating, I mean, we're spending a lot of time creating a brand that is really attracting in Paris. And it has been a really a great thing for us to build the team. We as well, super excited to be working with. I think that's been our most mostly differentiated approach. Here is around the locality of the engineering team for sure.

• 1:08:43Speaker 02

  Yeah, we're mostly based in San Francisco, so it's been a lot tougher. I'd say like, you know, like one, we, we, we hire more kind of like just software engineers as opposed to research engineers. And so like the, the folks who are getting a lot of the crazy salaries were probably not competing for them. That being said, like, you know, OpenAI and Anthropic and everyone is also hiring a bunch of software engineers as well. I think, you know, we're, we're a lot smaller than them. We're more of a startup. A lot of people want to work at a startup for a variety of reasons. And so that's been the main reason that someone would join us as opposed to one of the model ups.

• 1:09:16Mario Gabriele?

  Amazing. Well, I want to just ask one more AI question before we do a bit of our, our final wrap ups, which is more of a general one, but what are your rough sort of frameworks for when we can expect sort of true AGI, if you think we haven't hit it yet already? And, you know, ASI, you know, are you on the AI 2027 timeline? Are you more bullish, less bullish, more scared, less scared?

• 1:09:41Speaker 00

  Yeah, I don't really have a timeline. What I'm, what we've seen so far, if you should look back, is that investments in that ecosystem has been a leading indicator to the progress that has been made, which makes sense and is pretty obvious. I think looking at advancement, it's nowhere close to being stopped, so we can expect more progress. The pace of it, etc. is very hard to anticipate in any way. I mean, at the end of twenty twenty four, we were like, ah, things start plateauing and all of the sudden, uh, you've got a new paradigm that kind of emerge and and push back again or and push back the accelerator in terms of progress. So it's all very hard to anticipate. What is true is that it seems like the investment is still going crazy,

• 1:10:22Speaker 00

  which means that there is no uh uh no limits to uh to uh the kind of uh uh resources that will be invested in in the in making those models better. It is hard to believe at the same time that there is kind of a hidden uh like uh limits that supposedly would be here around before we reach even better capabilities. So I think it's mostly a question of uh of pace. To be honest, that's not something I'm spending too much time on, because I think we are, we, and I I probably assume that it's simple, Harrison, but we are pretty agnostic to the, or pretty uh uh edged in the, in both scenarios.

• 1:11:04Speaker 00

  I think the scenario is leading to something that is really superhuman and the machine take over all work. It will require some amount of time to transition to that, and there's going to be a lot of product to manage that transition. Eventually, everything's off anyway, so in a sense, you were pretty neutral. And if the technology kind of platos, I think the uh the uh uh deployment of that technology in the society uh will still take years and years and years. And so I think there's still also a lot of value to be constructed, to to be, to be, to be created around building the product that helps that deployment. So I think we're excited to be building in both world. Uh, one of the funky funny ways that motivated me to go back to building a a product was uh

• 1:11:49Speaker 00

  that, which is the last, last train before AGI. Uh, so it's kind of the last opportunity to be building a company. And so that was kind of, it's not the true motivation, but one of the uh fun uh reasons why I moved back from openAI and started building a startup. So uh, so we'll see. I think it's uh to me, um, I don't have a timeline and I I'm I'm I'm just amazed by how fast it's been progressing. And so I don't expect it to stop. And so I'm really wondering what we're going to be in two years. Don't know if it's going to be AI twenty twenty seven, but it's surely going to be uh uh crazy compared to where we are today.

• 1:12:24Mario Gabriele?

  Harrison, I saw you nodding along through through most of that. Does that sort of so how you see things more or less?

• 1:12:31Speaker 02

  Yeah, I think that's pretty spot on. Like, I don't spend a ton of time thinking about that either. I think there's, I think even, you know, even if the malls get really, really good, in order to make them impactful, you'll still want to integrate them in some way. And that has to happen somehow. And I think that's the type of work that that we focus a bunch on. I don't have any particular insights. And so I don't claim to, and I don't spend a ton of time thinking about it.

• 1:12:57Mario Gabriele?

  Well, let's move to two sort of wrap up questions, if that sounds good to you. Harrison, let's stay with you. If you had unlimited resources and no operational constraints, what is an experiment you would love to run?

• 1:13:08Speaker 02

  I don't know the exact experiment, but maybe two areas. One, one, just one general one, and then one more kind of like focused on this. I think memory is really, really interesting. I think memory for AI and agents is and personalization and learning, or whatever you want to call it. And so I don't know exactly what the experiment I would run is, but that's an area that I would absolutely kind of like explore. And then one that's like not related to AI at all, but like, how do I get the best sleep? Like, I just want to sleep really well. Like, I need like eight hours or I'm terrible the next day. And so like, what conditions set me best up for success there? Like, that's one that I personally would love to have an answer to.

• 1:13:44Mario Gabriele?

  Well, we're seeing some amazing things happening around like sequencing short sleepers and trying to figure out, you know, how to turn that into some kind of therapeutic where, you know, on four hours a night, maybe you can have the same level of productivity and energy or more. So who knows? Maybe, maybe we're just a few years away.

• 1:14:00Speaker 02

  That would be great.

• 1:14:01Mario Gabriele?

  Stan, what about you? What would your, your experiment be?

• 1:14:04Speaker 00

  At the end of the day, a lot of, and that's what we're doing, is just be about doing it faster. To some extent, I think at the end of the day, those models are still pretty smart. They do a lot of really impressive stuff. It's mostly that we don't, we don't have the pipes to connect them to the right actions and the right data. And so there's a lot of pipes missing. And so if you could send everybody, I mean, a very large team to do partnership with every platform up there and create all the pipes, et cetera, and see. How much and be then able to really figure out how much of work can be taken by those models? Because I think we don't see it quite clearly. Not only because there's the model capabilities that is a limitation.

• 1:14:47Speaker 00

  There is also the availability of the actions, availability of access to the data, which is never perfect. And so getting perfect there is purely an operational play. It's about building stuff or getting partnerships and stuff. And so that would be one of those. The other thing that is kind of a sidetrack is I'm super interested into answering the question of whether there is a maximum team size that exists for a given product, especially in the age of AI. I think many people do believe that there is an optimal team size for a given product, let's say. And that often the scale that some companies go into, to thousands of employees and stuff like that,

• 1:15:32Speaker 00

  is mostly people getting busy by just organizing themselves, in a sense. I think with agents being ambient around them and being used in many ways, this has become more and more true. And so if I could really quickly find out, you know, the scale of the team with really great talent at many different levels and quickly operate it to see how it feels, that would be something that I would be very interested to figure out. So if there is such maximum of efficiency at some point.

• 1:16:04Mario Gabriele?

  I love that. Yeah, where you hit the limit. Okay, this is a question I love to end with usually. If you had the power to assign a book to everyone on earth to read and understand, what book would you like to assign? And why don't we stick with you, Stan, and we can end with Harrison?

• 1:16:23Speaker 00

  Yeah, one book that I loved is, and I'm not sure I understood it completely.

• 1:16:31Speaker 00

  It's called, it's Greg Egan Permutation City, because I think it just sends you into thinking about the nature of consciousness in a way that is extremely interesting. And so that's one that comes to mind.

• 1:16:45Mario Gabriele?

  Is it a novel?

• 1:16:46Speaker 00

  It's a novel.

• 1:16:47Mario Gabriele?

  It's a novel, okay. Yeah, it sounded like it might be. Wow, that sounds really interesting. I'm gonna add that to my list. And what about you, Harrison?

• 1:16:54Speaker 02

  One of my favorite books is Range by David Epstein. And it actually, you know, it's great for the title of the podcast, The Generalist, because it's all about how generalists succeed in kind of like a specialist world. And I thought that was really interesting. And, you know, when they succeed and when it's also good to be a specialist and, you know, just how a lot of the great things in humanity have come from just generalists with range, connecting dots and putting things together. And so I really enjoyed that.

• 1:17:25Mario Gabriele?

  What a perfect coda for us here. Thank you both so much for taking the time. I learned so much and yeah, really enjoyed chatting with you both.

• 1:17:37Mario Gabriele?

  That's it. Thank you for listening to this episode of The Generalist Podcast. Please subscribe on Apple Podcasts, Spotify, or your preferred podcast app. Ratings and reviews help others discover these discussions. So if you enjoyed the conversation, I'd be grateful if you could take a moment to leave one. For all past episodes and more, visit us at thegeneralist.substack.com. See you next time as we continue to explore the future.