Skip to content

Podcast

The Fastest Path to Super Intelligence

Y Combinator Startup Podcast

Source ↗ ← All highlights
  • Recursive Self Improvement Without Retraining
    • Recursive self-improvement can be built on top of existing LLMs instead of training new models from scratch.
    • Ian Fischer says Poetic uses a meta-system that sits on frontier models as “stilts” to make them smarter without expensive retraining. Transcript: Ian Fischer At Poetic, what we’re building is a recursively self-improving system. And so recursive self-improvement is this 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. 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 hundreds of millions of dollars and takes months of effort. Unknown Speaker And then Anthropic or OpenAI will come along and just eat your lunch in the next model release. Ian Fischer Right, right. And of course, Anthropic and OpenAI and Google, they’re exploring recursive self-improvement, but typically at that level of having to train a new model for every step of self-improvement That they do. Unknown Speaker I mean, that seems like actually the defining thing that a startup really, really wants. I know that I want to take advantage of whatever the next model is, but the second you’re in fine-tuning land, I’m spending millions to hundreds of millions of dollars. And then guess what? I just lit it on fire because the next version of the frontier model comes out and I’ll never catch up. Whereas 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. Yeah. Ian Fischer 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. (Time 0:01:07)
  • Model Agnostic Harnesses Beat The Bitter Lesson
    • Poetic’s harnesses are model-agnostic and remain compatible when newer, stronger base models arrive.
    • The same harness gives an immediate performance bump on new models and can be further optimized for them. Transcript: Ian Fischer You probably would have said, okay, we’re going to first collect a large data set, you know, like tens of thousands of examples for our particular problem that we’re working on. And we’re going to fine tune, you know, the best model we can put, 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 fine-tuned on top of. But by then, a new model has come out, and it’s better than the thing that you fine-tuned. You fine-tuned three years ago on top of GPT-3 or whatever, and then GPT-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 in some cases, the latter. With Poetic, what we end up giving you is a, you know, people are calling these things harnesses now, but, you know, or a Gentic 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. (Time 0:03:32)
  • Poetic Leapfrogged Gemini On ARC-AGI
    • Poetic rapidly outperformed Gemini 3 Deep Think on ARC-AGI V2 shortly after Gemini’s release.
    • Ian Fischer notes their run cost about half as much by using Gemini 3 Pro and achieved a nine percentage point lead. Transcript: Ian Fischer Arc AGI V2 was kind of us coming out of letting people know that we could tackle these really hard problems. And in particular, we wanted to show that our system could generate these, what we call our system like the poetic meta system, can generate reasoning systems that are highly effective. Gemini 3 Deep Think had just come out, and they were really quite dramatically at the top of the leaderboard at 45%. And two days later, we released our results where we were showing that we could get a lot higher than that. Unknown Speaker So they come out with soda, and then you come in right above them every single time, which is 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. Ian Fischer Yeah. So the interesting thing is that we were half the cost of Gemini 3 Deep Think because we were building on top of Gemini 3 Pro, which is a much cheaper model. But we still got, in the end, a nine percentage point improvement on the official verification. (Time 0:05:19)
  • Top Score On Humanities Last Exam With A Small Team
    • Poetic achieved 55% on Humanities Last Exam, beating Anthropic’s recent 53.1% with Claude Opus 4.6.
    • Ian Fischer reveals their optimization cost under $100k and the team is seven people of researchers and engineers. Transcript: Ian Fischer Humanities 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 Opus 4.6. They got 53.1% and we got 55% on it. Other Speaker And one thing that Humanity Last Exempt 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? Ian Fischer We didn’t publish any cost for this, but I can say that the optimization costs us less than 100k. (Time 0:06:44)
  • Meta System Produces Full Reasoning Harnesses
    • Poetic’s meta-system outputs code, prompts, and data that form reasoning harnesses solving tasks too hard for base models alone.
    • The system can also optimize existing agents’ prompts, reasoning strategies, or specific components automatically. Transcript: Ian Fischer These harnesses, they are code, prompts, data, built on top of one or more language models. And so this is something that in principle you can build by hand or with like cloud code or whatever. It takes a lot of work to do these, to have all the insights 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, where a hard problem is something that if you gave it to GPT-5 it would struggle to give you a reliable, robust result, 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 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, 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. (Time 0:08:43)
  • Shifting S Curves Accelerate Capabilities
    • Each underlying model family and Poetic’s meta-system each have their own S-curve of capability that shift as both improve.
    • Ian Fischer suggests stacking improvements could push performance toward AGI or superintelligence over time. Transcript: Ian Fischer 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’ll find that the S-curve that you’re dealing with keeps shifting higher and higher until ultimately Either you saturate or like… Reach AGI? Yeah, reach AGI, reach super intelligences. Yeah. Unknown Speaker Given its stilts, you might like hit the ceiling first then. That’s the goal, (Time 0:10:47)
  • Outsourcing Prompt Engineering To The Meta System
    • Poetic delegates dataset understanding and failure-mode discovery to the AI, rather than humans manually inspecting data.
    • The meta-system decides when to add context, generate examples, or change strategies to improve robustness. Transcript: Ian Fischer Instead, we’re letting the poetic meta system look at that data. And so like the meta system, you know, if it, if it thinks that it needs to put more things into context, do more context stuffing or whatever, it’ll do that. If it needs to like generate a bunch of examples 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 ArcGIS in that I think you can read those and say, well, that’s not what a human would have written pretty Clearly. And there’s some unexpected stuff. And 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 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. (Time 0:11:58)
  • Reasoning Strategies Multiply Performance
    • Manual prompt tuning yielded small gains (5%) on a hard task, but adding programmatic reasoning strategies raised performance to 95%.
    • Ian Fischer contrasts prompt tweaks with code-based reasoning harnesses as the major multiplier. Transcript: Ian Fischer And what we saw there is that 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 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%. (Time 0:13:55)
  • Bring Hard Agent Problems To Poetic Early Access
    • If you have a hard problem and existing agents stall, sign up for Poetic early access to get their optimization help.
    • Poetic invites startups to describe difficult, reliability-limited problems so they can be contacted for collaboration. Transcript: Ian Fischer 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 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 what it is that you’re working on and we’ll reach out. (Time 0:14:57)
  • Experiment Daily With AI To Learn Quickly
    • Try AI projects constantly and push the boundaries by building small experiments every day.
    • Ian Fischer recounts building an iPhone app in a weekend with GPT-5 and urges founders to just try things. Transcript: Ian Fischer Like, how should they think about that? 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 with AI. Always try to push yourself to find the boundaries of what they’re capable of and build the things that you want to build. Even for me, 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. So fast. Yeah, it’s so fast and so easy. And that was, you know, that was an, 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. (Time 0:18:26)