The Fastest Path To Super Intelligence

Poetiq is a new startup founded by former DeepMind researchers that recently achieved a major jump on the ARC-AGI and Humanity's Last Exam benchmark by layering a recursive self-improvement system on top of existing models. In this episode of Lightcone, Poetiq's Founder & CEO Ian Fischer joined us to discuss how small teams can build “reasoning harnesses” that outperform base models, what that means for startups and why automating prompt engineering may be one of the most powerful levers in AI today. Chapters:00:00 – Intro00:40 – What Is Poetiq?01:07 – Recursive Self-Improvement Explained02:07 – The Fine-Tuning Trap02:59 – “Stilts” for LLMs03:14 – Recursive Self-Improvement vs. Fine-Tuning05:05 – Taking the Top Spot on ARC-AGI06:37 – Beating Claude on Humanity’s Last Exam08:40 – How the Meta-System Works10:26 – Beyond RL: A New S-Curve11:32 – Automating Prompt Engineering13:37 – From 5% to 95% Performance14:50 – Early Access & Putting Your Agent on Stilts16:17 – From YC Founder to DeepMind Researcher18:29 – Advice for Engineers in the AI EraApply to Y Combinator: https://www.ycombinator.com/applyWork at a startup: https://www.ycombinator.com/jobs

• 0:00Speaker 00

  The world is changing so quickly. This is probably a little bit obvious, but you should just try things and like every day, do something with AI. Last summer, I took a weekend and used GPT-5 to help me build an iPhone app. I hadn't done that in a decade.

• 0:17Speaker 01

  So fast.

• 0:18Speaker 00

  Yeah, it's so fast and so easy. And that was, you know, an age ago. That was like eight months ago. Now it's even faster and easier. Don't limit yourself. Like anything that you imagine, you should just try to use AI and see how far you can get with it. And you'll be, you know, making the world better.

• 0:40Speaker 01

  Welcome to another episode of The Light Cone. Ian Fischer is the co-founder and co-CEO of Poetic, which is building recursively self-improving AI reasoning harnesses for LLMs. Previously, he spent a decade as a researcher at Google DeepMind and founded a mobile dev tools company through YC years ago. Welcome, Ian.

• 1:00Speaker 00

  Thank you. I'm so happy to be here.

• 1:01Speaker 01

  What is Poetic? How's it different than RL? You know, how's it different than context engineering?

• 1:07Speaker 00

  At Poetic, what we're building is a recursively self-improving system. And so recursive self-improvement is this, you know, kind of the holy grail of AI, where the AI is making itself smarter. The core insight that we had is that we could do recursive self-improvement far faster and cheaper than all of the other ways that people had been proposing to do this. And so obviously, I can't go into details about what that is, what our particular approach is. But most of the approaches out there involve, you know, they require you to train a new LLM from scratch. And training LLMs from scratch costs, you know, hundreds of millions of dollars and takes months of effort. And so the-

• 1:49Speaker 01

  And then Anthropic or OpenAI will come along and just eat your lunch in the next model release.

• 1:53Speaker 00

  Right, right. And, you know, of course Anthropic and OpenAI and Google, they're exploring recursive self-improvement, but typically at that level of having the, you know, having to train a new model for every step of self-improvement that they do.

• 2:07Speaker 01

  I mean, that seems like actually the like defining thing that a startup really, really wants. Like, I know that I want to take advantage of whatever the next model is, but the second year in fine-tuning land- I'm spending millions to hundreds of millions of dollars. And then guess what? Like, I just lit it on fire because the next version of the frontier model comes out, and I'll never catch up. Whereas like, working with your systems means that I will always have the thing that is better than the thing that's out of box. And that's sort of like the holy grail.

• 2:38Speaker 00

  Yeah, we think that this is incredibly valuable to anybody who's building on top of large language models. And we don't view the frontier models as competitors. They're the ones that were using the stilts, building stilts to stand on top of. But if we didn't have that foundational layer, then Poetic couldn't exist.

• 2:59Speaker 01

  Yeah, I mean, being the smartest model, it's a game of inches, actually. And so those inches matter a lot.

• 3:05Speaker 00

  Right, right.

• 3:06Speaker 01

  How do we actually get started? I mean, you've built something that basically any startup could use, that it's sort of like stilts, really.

• 3:14Speaker 00

  We have built a system that can automatically generate systems for your particular problem that will always outperform the underlying language models. And without the massive expense, as you're saying about the bitter lesson, where what would you have done without Poetic? You probably would have said, okay, we're going to first collect a large data set, like tens of thousands of examples for our particular problem that we're working on. And we're going to fine tune the best model we can get our hands on. Maybe that's one of the frontier models, or maybe it's an open weights model. It doesn't particularly matter. You're going to spend a lot of money on that fine tuning. The compute is so expensive. And then at the end of it, you have something that works better than the thing that you

• 3:59Speaker 00

  fine tuned on top of. But by then, a new model's come out. And it's better than the thing that you fine tuned. You fine tuned like three years ago on top of GPT 3.5 or whatever. And then GPT 4.4 comes out, and it just blows you out of the water. And so are you going to do that again? Or are you going to go out of business? And like, in some cases, the latter. With Poetic, what we end up giving you is a, people are calling these things harnesses now, but we're an agentic system. Or whatever you want to call it that sits on top of one or more language models. And it just performs better than them. And when the new model comes out, that same harness is perfectly compatible with it.

• 4:41Speaker 00

  And you don't need to change anything to get the, you know, an even bigger performance bump. Additionally, we can, you know, continue to optimize for this new model. Whatever the new model is that you want to use and, you know, make it even better. But you you don't lose out on, you know, hundreds of millions of dollars. In fact, we do this so much more cheaply than fine tuning would cost as well.

• 5:05Speaker 01

  And you've done this actually a bunch of times, right? Like, I remember when you first came out with your paper in December of last year. You shot to the top of Arc AGI V2. And then you've done this a bunch of times for other benchmarks, too. What, you know, what was that like?

• 5:19Speaker 00

  Arc AGI V2 was, this was kind of, you know, us coming out of stealth, letting people know that we could tackle these really hard problems. And in particular, you know, we wanted to show that our system could generate these, what we call, you know, we call our system like the poetic meta system. Can generate reasoning systems that are highly effective. Gemini three deep think had just come out and they were, you know, really quite dramatically at the top of the leaderboard at forty five percent. And two days later, we released our results where we were showing that we could get a lot higher than that.

• 5:59Speaker 01

  So they come out with soda and then you come in right above them every single time.

• 6:04Speaker 00

  Yeah.

• 6:04Speaker 01

  Like wild to see, honestly. That's what it's like to have stilts, you know, like whatever model comes out, you can be taller than that one with poetic, which is like, that's so awesome.

• 6:13Speaker 00

  Yeah, so the interesting thing is that we were half the cost of Gemini three deep think. Because we were building on top of Gemini three pro, which is a much cheaper model. But we still got in the end a nine percentage point improvement on the official verification. So they were at forty five percent and we were and like seventy something dollars and we were at fifty four percent and thirty two dollars per problem.

• 6:37Speaker 02

  So recently, you guys just announced some incredible results for humanities last exam. Can you tell us more about those?

• 6:44Speaker 00

  Humanity's Last Exam is a set of 2,500 really, really hard questions written by experts in many different domains. They're meant to be challenging even for PhDs in those fields. AI hasn't passed it yet, but we got to 55%, which is almost two percentage points higher than the previous state of the art, which came out just last week from Anthropic with Claude 4.6. They got 53.1% and we got 55% on it.

• 7:19Speaker 02

  And one thing that Humanity Last Exam doesn't publish is the cost of getting those results. In your case, this run was done with less than around six figure? How much was it?

• 7:31Speaker 00

  We didn't publish any cost for this, but I can say that the optimization costs us less than 100K. Yeah.

• 7:38Speaker 02

  Which is impressive because each of these big foundation models trained runs are in the hundreds of millions of dollars. And you guys, as a company, you're only seven people?

• 7:49Speaker 00

  That's right. Yeah. Seven, seven research scientists and research engineers. Yeah.

• 7:53Speaker 02

  That's impressive. And I think the thing that's very interesting about your approach is sort of taking a very scientific approach to the emergent behaviors that a lot of the best founders are doing with models. I think a lot of founders that get very good results for agents, they treat the underlying model as a common layer that you can switch in between. And there's a certain task, for example, for GPT 5.2, like very hard to verify bugs, gets sent to that versus architecture that gets sent to Claude 4.6. But you're kind of doing this automatically instead of having a human conducting is very impressive. I think there's something more special going on underneath. Can you tell us a bit about How it works.

• 8:37Speaker 01

  Yeah, it sounds magical. So what can you tell us?

• 8:40Speaker 00

  Right. So you're getting at a core, a really core thing. You know, these harnesses, they are code, prompts, data, you know, built on top of one or more language models, right? And so this is something that in principle you can build by hand or with like Claude code or whatever. But in in practice, it takes a lot of work to do these, to, you know, have all the insights to make this, to make these work well. And so the core technology that we've developed at Poetic is recursive self-improvement. So we have a recursively self-improving system, which we call the Poetic meta system. The output of that system is systems that solve hard problems,

• 9:22Speaker 00

  where a hard problem is, you know, something that if you gave it to GPT-5-2, it would struggle to give you a reliable, robust result, you know, just to use an example. So this is a very big advantage for us. We can generate these systems in a much more automated manner, which means that we can do it much more quickly and much more cheaply than if you hired a team yourself to try to make your own, you know, your own agent to solve your particular task. But not only that, since, you know, this is really an automated optimization process, if you already have done that work, you, you know, you're a startup that's like going after a particular vertical and you've put together, you know, you think you understand your problem pretty well. You've put together your agent and you, you know,

• 10:06Speaker 00

  maybe it's working pretty well, but you know you can get something better or you really need something better, then you can bring that to us and we can optimize that entire agent or pieces of that agent. We could optimize just the prompts, just the reasoning strategies. There's a lot of different things that we can do depending on your particular needs.

• 10:26Speaker 02

  It sounds like this is a completely different paradigm than RL, because we went through the S curve of regular pre-training. RL with when OpenAI release O1. And now this feels like a new one. It sounds special. It sounds it rhymes a lot with RNNs, which is a whole different paradigm than RL, right?

• 10:47Speaker 00

  It's going to depend on the particular task, the particular type of problem that we're going after, that we're trying to solve, and the underlying models that we're working with. But effectively, you could say like each model or each set of models that we're working with will have their own S curve. The poetic system, the poetic meta system itself, is also going to have its own S curve. And so as a poetic meta system gets better and as the underlying models get better, you you'll find that the you know, the S curve that you're dealing with keeps shifting higher and higher until ultimately either you saturate or like.

• 11:20Speaker 02

  Reach AGI?

• 11:21Speaker 00

  Yeah, reach AGI, reach super intelligences. Yeah.

• 11:24Speaker 01

  Given it's stilts, you might like hit the ceiling first then.

• 11:27Speaker 00

  That's the goal, right?

• 11:28Speaker 01

  Yeah.

• 11:28Speaker 00

  You want to hit the ceiling first with poetic.

• 11:31Speaker 01

  I think a lot of startups that we work with. And then in my spare time, I, you know, do a bunch of context engineering. And then the thing is, we're sort of like tuning it, tuning evals, tuning like we're context stuffing ourselves. What does that even feel like to have a, you know, recursively self improving version of like prompt engineering and context engineering?

• 11:52Speaker 00

  We don't spend a lot of time looking at the particular data that we're working with. Instead, we're letting the poetic meta system look at that data. And and so like the meta system, you know, if it if it thinks that it needs to put more things into context, it do more context stuffing or whatever, it will it will do that. If it needs to like generate a bunch of examples

• 12:16Speaker 00

  to get the get better performance, it'll do that for you, right? It was pretty interesting to look at the prompt outputs in particular, I'd say for Arc AGI. In that, you know, I think you can read those and say, well, that's not what a human would have written pretty clearly. And and it's, you know, there's some unexpected stuff. And, you know, it made some really simple examples. And one of the examples is actually wrong. But we didn't change it. We're like, well, this is, you know, this is the thing that it output. We'll just leave it be. You know, we don't want to go in and monkey around with things. And so historically in machine learning, you always, you know, it's like the the rule was you have to know your data set really well.

• 12:59Speaker 00

  But now we're kind of outsourcing that to the AI itself, where the AI is, it's the AI's job to understand the data set and figure out where are the failure modes? And where are the kind of robust reasoning strategies that the model, that the agent could use to get better performance?

• 13:19Speaker 01

  How much of it is like much, the output is much better prompts? And then how much of it is like the harness itself, context stuffing, or summarizing in the right way, or reranking in the right way? So that like you have some number of like mega LLM calls? And then how do you get the most out of each of those calls?

• 13:37Speaker 00

  Yeah, and so that definitely varies per problem. But what we've seen, in fact, our last paper at DeepMind was not doing this recursive self-improving stuff, but we were, we were showing that you could build these harnesses manually to solve really hard problems. And what we saw there is that, you know, we manually optimized the prompts really hard for these very hard problems. And that got us a little bit of the way. In this particular case, you know, the hardest, the hardest task we were working on, we got like to 5% performance with Gemini 1.5 Flash. This was a while ago. And then when we added on the reasoning strategies, we went from 5% to 95%.

• 14:22Speaker 00

  And so this is typically what we see, you know, like everybody's out there kind of doing some amount, I wouldn't say everybody, but many people are out there kind of doing some amount of automated prompt optimization. You know, JEPa is this very popular paper. Everybody's kind of re-implementing that. That will get you some performance improvements. But it's very far from everything that you can get if you actually think about these reasoning strategies that are really going to be written in code rather than in just better prompts.

• 14:50Speaker 03

  So if startups want to use Poetic to put their agent on stilts, what should they do?

• 14:57Speaker 00

  Yeah, so right now, we haven't released anything yet. But if you go to poetic.ai, there is a button you can click to get sign up for early access. And if you're a startup or a company who has a really hard problem and you've tried everything that you can to make it reliable and robust, and you just can't get all the way there, you need something more, then let us know. We're looking for problems like that. So just tell us, tell us what it is that you're working on and we'll reach out. You'll be the first to know when we're, when we're ready to work with you.

• 15:30Speaker 01

  I mean, if you're at the top of humanity's last exam, then I mean, that's, that's pretty big. So you're all, you're already all the way out there at soda. And then I guess the stilts basically let any agentic company become soda.

• 15:44Speaker 00

  That's the idea. Yeah, yeah. And, you know, we view the RKGI results and the humanities last exam results as showing kind of two different capabilities that we have. We can really improve your reasoning and we can really improve deep knowledge extraction from these models.

• 16:00Speaker 01

  And then you're just totally vaccinated against the bitter lesson.

• 16:03Speaker 00

  Exactly.

• 16:03Speaker 01

  YC's next batch is now taking applications. Got a startup in you? Apply at ycombinator.com/apply. It's never too early and filling out the app will level up your idea. Okay, back to the video.

• 16:17Speaker 04

  A slight change, change of topic, but something I was curious about. You arrived at Google over a decade ago when they acquired your first YC startup, a portable. A portable was it's porting mobile apps cross-platform, right? Like Android or whatever. It's quite different to recursive self-improving AGI. How did you make that leap? What happened once you got to Google? What made you think that you maybe wanted to shift out and do something different? And just would love to hear that story.

• 16:46Speaker 00

  The acquisition was this amazing opportunity to reflect on what I really wanted to be doing next, right? Like, Google was in the, you know, itself is a place where you can do so many different things. So I spent some time thinking about where, where I wanted to go next in, in my journey. I realized that the problems that I was most excited about were really actually AI and, and robotics.

• 17:21Speaker 00

  And it's the best people in the world. Many of them in those fields were at Google at the time. And so I went and talked to them. They let me come join, you know, a new AI robotics team in Google Research, which was this amazing opportunity for me since that wasn't my background. My background was like computer security and then this cross-platform mobile, you know, it's systems building stuff. I was able to join this team and I'll tell you the truth that I very quickly realized that hardware is hard and I didn't really want to be doing robotics. This is more aspirational at that moment, but I was really passionate about machine learning.

• 18:04Speaker 00

  So I just made a very hard switch into just doing machine learning research and did that for, you know, about a decade at Google and then Google and then DeepMind.

• 18:16Speaker 04

  What's maybe some advice that you have today for engineers who want to get into more of the AI side, probably the applied AI and build startups around AI? Like, how should they think about that?

• 18:28Speaker 00

  You know, the world is changing so quickly. This is probably a little bit obvious, but you should just try things and like every day do something, do something with AI. Always try to push yourself to find the boundaries of what they're capable of and build the things that you that you want to build, right? Even even for me, you know, last summer I took a weekend and used GPT-5 to help me build an iPhone app. I hadn't I hadn't done that in a decade.

• 19:02Speaker 04

  So fast.

• 19:03Speaker 00

  Yeah, it's so fast and so easy. And that was, you know, that was, you know, an age ago. That was like eight months ago. Now it's even faster and easier. Don't limit yourself. Like, anything that you imagine, you should just try to use AI and see how far you can get with it and you'll be, you know, making the world better.

• 19:19Speaker 01

  That's all we have time for today, but Ian, thank you so much for giving us all stilts. We can't wait to use it at YC. I can't wait to use it for Gary's List. I mean, there's just so much to do, so.

• 19:31Speaker 00

  Yeah, thank you for having me. This was a lot of fun.