Postdoc at Danish Cancer Institute. Interested in cancer genomics and applying LLMs in science. Follow me for news and repos I find online about AI and science.
Copenhagen, Denmark
Joined December 2016
- Tweets268
- Following680
- Followers185
- Likes1.2K
Pinned Tweet
Yesterday I finally set up one of the largest, open-source LLM models in the world, GLM-5.2, on Gefion, the Danish supercomputer, to find new synthetic lethal relationships in cancer. It's going strong at around 5 tokens/second (I have to optimize the inference a bit...), evaluating a new gene pair every ~15 mins.
Qualitatively, the reasoning capability of the model already looks better than that of Qwen 32B-instruct, but requires 2 full nodes, 2x8 H100 GPU-s to run it, and it's not quite trivial to set up the inter-node sharding.
See comments for our manuscript, now in peer review.
I'm experimenting with mapping out all the ongoing community efforts to improve, validate, or extend the mathematical results OpenAI released earlier this week. Check it out below!
math-frontier-atlas.netlify.…
The app is open source if you'd like to host it or contribute to it: github.com/Paureel/math-fron…
The app is open source if you'd like to host it or contribute to it: github.com/Paureel/math-fron…
Thanks everyone for the feedback, the integer multiplication tracker now has a zoom feature, so hopefully you will be able to see the most recent progress better! Let me know in the comments if you have other ideas to add!
beyond-n-log-n.netlify.app
The tracker itself is also open-source, feel free to customize it to other community efforts: github.com/Paureel/beyond-n-…
I've just made a live tracker for the community effort for optimizing the "Integer multiplication below n log n" mathematical problem, so everybody can see the progress on it: beyond-n-log-n.netlify.app
Are you interested in modifying this, or building new trackers for the different mathematical problems? I open-sourced the tracker here: github.com/Paureel/beyond-n-…
Aurel Prosz retweeted
Went to bed last night, and sorry to report only a bit of incremental progress on my side.
*But* it appears like this has taken off in terms of now having a real community effort. Really inspiring to see what looks like major progress from some of you.
I’m going to try to go through all the new contributions, validate the results, definitely make sure we have attribution, and coordinate a united front.
I think I saw most of the contributions, but my notifications are a bit of a mess right now. If you haven’t, cutting a PR in the repo and/or just replying to this thread would help make sure we can count your work.
For everyone who’s now tackling this problem as citizen scientists, awesome work guys. Really inspiring stuff. Let’s see how far we can push this
We are publishing an update to OpenAI problem #109 (integer multiplication) with further tightening.
κ = 2⁻³⁴ (tightened from κ = 2⁻¹⁸²)
The exact witness is 8.3 × 10⁻¹¹, a roughly 48 million fold improvement over the previous result and a 2¹⁴⁸ fold improvement over original OAI result.
The improvement came from removing the spacing penalty behind the quadratic bottleneck. This was done by moving compact control bits instead of entire windows.
If you want to understand the current breakthrough in AI about resolving the Navier-Stokes Millenium Prize problem, check out this interactive textbook I made with GPT-6 Astra. You can select a explanation level from kindergartener to math professor.
explain-navier-stokes.netlif…
If anybody is interested in the technical details:
- I downloaded the pdf containing the proof from OpenAI website
- Used GPT-6 PRO to create a detailed prompt for Codex
- Used Coded with gpt-6-astra extra-high mode to generate the app
- Deployed on Netlify
Build time ~2 hours end-to-end.
Keep in mind, this is an educational tool, LLMs can hallucinate, be mindful when interpreting the results, and all the usual stuff.
One of the reasons I'm optimistic about AI progress is that it unlocks our creativity in new ways to communicate our research: over the weekend I built a working 3D video game for fun that lets you control a miniature spaceship inside a cancer cell, grabbing proteins and letting the original algorithm used in our synthetic lethality project calculate the probability of those being a good candidate for further evaluation. Funnily enough, this is a super inefficient way to run the algorithm, but it's a good way to show what our research is about to people who are not familiar with the field, instead of just showing a graph or a list of numbers. And if you are lucky, you can actually pick the right combination, but good luck, there are around 400 million to choose from.
Some technical details:
- Okay, to be fair, it was codex from OpenAI and gpt 5.6-sol-xhigh that built the game, I just supplied the vision, the general direction and encouraged the AI on the way. I have zero expertise in game or 3D asset development, but this is also kind of the point.
- All assets, animation, textures were generated by codex. The whole build time was around 4 hours, and utilized ~30 subagents
- I used @mattshumer_ 's Gauntlet Loop for sub agent orchestration, link to it is in the comments
- The game currently does not use real physical simulations on how the molecules move in the cell, very limited amount of genes/proteins to pick, etc, and is illustrative at this point. If there is interest I'll keep working on it and open-source it so anybody can try it out.
What do you think about this approach to communicate research? What will you build?
I built a debugger and visualization platform for deep and wide recursive agent swarms, so you can monitor how your harness evolves in real time.
RLM Agent Observatory shows topology, execution history, per-agent output, causal traces, and learned knowledge in one live view.
Focusing on Prime Agent today, but more frameworks coming next. 🧵
Knowledge belongs to the run.
Memories, skills, prompt notes, and refinements keep their provenance. Save or load the complete run and knowledge together. It natively matches the Prime Agent's harness state, so the project self-learns.
Open source: github.com/Paureel/rlm-agent…
What would you inspect first? I'm using it on launching agent swarms to find novel synthetic lethal targets in cancer, but comment your ideas below.
It's interesting to see that so far our reality follows quite closely the AI 2027 predictions (ai-2027.com/), started a year ago. If you haven't read it already, give it a go. I wonder if it eventually diverges somewhere significantly, since the predictions for the next 6 months are quite out of a sci-fi book. By early 2027, AI 2027 expects AI to move from assisting research to materially accelerating the development of better AI.
