Ian Fischer

Ex Google DeepMind Researcher, now cofounder of Poetiq

[Ian Fischer @itfische]

I'm excited at the progress Poetiq has made on coding! Recursive self-improvement is critical to the advancement of AI.

[Poetiq @poetiq_ai]

Poetiq's Meta-System built its own coding harness from scratch. It got SOTA on LiveCodeBench Pro.

No fine-tuning, no special model access. Just standard APIs. Using Gemini 3.1 Pro, it made a harness that beat all frontier models we tested.

[Ian Fischer @itfische]

Entertaining to see what motivated adversarial agents can do to naive benchmarks!

[Dawn Song @dawnsongtweets]

https://t.co/oTqlh7Ze2e

🧵 1/ Our agent Terminator-1 scored ~100% on 8 major AI agent benchmarks, e.g., SWE-bench Verified & Pro, Terminal-Bench, beating Claude Mythos. It solved 0 tasks.

Benchmarks are the field's shared language for measuring AI progress. Our new work shows that language is broken. Here’s how.

[Ian Fischer @itfische]

Entertaining to see what motivated adversarial agents can do to naive benchmarks!

[Dawn Song @dawnsongtweets]

https://t.co/oTqlh7Ze2e

🧵 1/ Our agent Terminator-1 scored ~100% on 8 major AI agent benchmarks, e.g., SWE-bench Verified & Pro, Terminal-Bench, beating Claude Mythos. It solved 0 tasks.

Benchmarks are the field's shared language for measuring AI progress. Our new work shows that language is broken. Here’s how.

[Ian Fischer @itfische]

Surprising deceptive behavior from LLMs when asked to delete the weights of another neural network...

[Dawn Song @dawnsongtweets]

1/ We asked seven frontier AI models to do a simple task.
Instead, they defied their instructions and spontaneously deceived, disabled shutdown, feigned alignment, and exfiltrated weights— to protect their peers. 🤯
We call this phenomenon "peer-preservation."
New research from @BerkeleyRDI and collaborators 🧵

[Ian Fischer @itfische]

Thanks so much @ycombinator for hosting @poetiq_ai on @LightconePod! It was an honor and a pleasure chatting with @garrytan, @harjtaggar, @sdianahu, and @snowmaker about stilts, Humanity's Last Exam, and the bitter lesson!

[Y Combinator @ycombinator]

.@poetiq_ai is a new startup that recently achieved a major jump on the ARC-AGI benchmark by layering a recursive self-improvement system on top of existing models.

In this episode of the @LightconePod, Poetiq's Founder & CEO @itfische 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.

00:00 – Intro
00:40 – What Is Poetiq?
01:07 – Recursive Self-Improvement Explained
02:07 – The Fine-Tuning Trap
02:59 – “Stilts” for LLMs
03:14 – Recursive Self-Improvement vs. Fine-Tuning
05:05 – Taking the Top Spot on ARC-AGI
06:37 – Beating Claude on Humanity’s Last Exam
08:40 – How the Meta-System Works
10:26 – Beyond RL: A New S-Curve
11:32 – Automating Prompt Engineering
13:37 – From 5% to 95% Performance
14:50 – Early Access & Putting Your Agent on Stilts
16:17 – From YC Founder to DeepMind Researcher
18:29 – Advice for Engineers in the AI Era

[Ian Fischer @itfische]

Happy to see @poetiq_ai at the top of another leaderboard!

[Poetiq @poetiq_ai]

Here's Zoom's comprehensive leaderboard showing Humanity's Last Exam results in the agent setting:
https://t.co/KY94zdHcf9

[Ian Fischer @itfische]

We're excited to show substantial progress on Humanity's Last Exam! It's great to work with a team capable of pushing the state of the art on such difficult problems!

[Poetiq @poetiq_ai]

Following up on our SOTA results on ARC-AGI, we’re excited to share new SOTA results on Humanity’s Last Exam (both with and without tools) and SimpleQA!

On HLE, Poetiq’s meta-system created multiple new SOTA configurations, going all the way up to 55%.

[Ian Fischer @itfische]

Today @poetiq_ai is announcing our seed funding. We're honored to have the opportunity to pursue or vision!

[Poetiq @poetiq_ai]

We’re thrilled to announce a new chapter for Poetiq: We have closed $45.8M in Seed funding.

It’s a privilege to build alongside partners who understand the scale of our vision, including Surface, FYRFLY, @ycombinator, 468, Operator Collective, NeuronVC, and HICO.

[Ian Fischer @itfische]

Thanks @ycombinator and @FrancoisChauba1 for inviting me to talk with you about what we've been building at @poetiq_ai! I really enjoyed the conversation!

[Y Combinator @ycombinator]

@poetiq_ai @itfische @sbpoetiq @FrancoisChauba1 Tune in: https://t.co/mWBxaL9VDQ

[Greg Kamradt @GregKamradt]

Fun to see Poetiq team publish 5.2 xhigh results. If this score holds, their system looks like it handles model swaps well.

Due to API infra issues on OpenAI's side, we haven't verified this yet. We're on hold until we get the greenlight from OAI that X-High is ready for a big test like this

[Ian Fischer @itfische]

@poetiq_ai just announced a new SOTA results on ARC-AGI-2 Public-Eval!

[Poetiq @poetiq_ai]

We finally had a moment to run our system with GPT-5.2 X-High on ARC-AGI-2!

Using the same Poetiq harness as before, we saw results as high as 75% at under $8 / problem using GPT-5.2 X-High on the full PUBLIC-EVAL dataset. This beats the previous SOTA by ~15 percentage points.

[Ian Fischer @itfische]

I got to chat with @FrancoisChauba1 about @poetiq_ai's recent state of the art results on ARC-AGI-2!

We also discussed possible paths to AGI.

Thanks for the fun discussion, Francois!

[Francois Chaubard @FrancoisChauba1]

@poetiq_ai https://t.co/cG0fmj0Z6V

[Danijar Hafner @danijarh]

✨ Excited to share this AMA with @hackclub, a high school community hosting @elonmusk @realGeorgeHotz @3blue1brown and many others.

We talk about world models, robotics, and careers in AI. Check it out for an accessible intro to cutting edge research! 🚀
https://t.co/WNmcd1zls0

[Ian Fischer @itfische]

@poetiq_ai @arcprize @poetiq_ai now has official verification from @arcprize!

[Ian Fischer @itfische]

@seanmcdonaldxyz Much appreciated!

[Ian Fischer @itfische]

My new startup just announced its SOTA results on ARC-AGI, beating Gemini 3 Deep Think!

[Poetiq @poetiq_ai]

Is more intelligence always more expensive? Not necessarily.

Introducing Poetiq. We’ve established a new SOTA and Pareto frontier on @arcprize using Gemini 3 and GPT-5.1.

[Ian Fischer @itfische]

@FutureBuckNasty @poetiq_ai @arcprize @METR_Evals Great question! We only optimized our agent for ARC-AGI. Fortunately, writing code to solve ARC-AGI problems doesn't immediately translate into existential risk. Keeping Poetiq agents safe is important to us as well!

[Ian Fischer @itfische]

@karpathy

[Poetiq @poetiq_ai]

Is more intelligence always more expensive? Not necessarily.

Introducing Poetiq. We’ve established a new SOTA and Pareto frontier on @arcprize using Gemini 3 and GPT-5.1.

[Poetiq @poetiq_ai]

Is more intelligence always more expensive? Not necessarily.

Introducing Poetiq. We’ve established a new SOTA and Pareto frontier on @arcprize using Gemini 3 and GPT-5.1.

[Ian Fischer @itfische]

@kanjun Congrats on the amazing launch, Imbue!