Our next-generation AI systems are helping scientists to tackle some of the world's most pressing challenges.| Google DeepMind
So the big news this week is that o3, OpenAI’s new language model, got 25% on FrontierMath. Let’s start by explaining what this means.| Xena
Import AI publishes first on Substack – subscribe here. Cambridge researchers show how to use distributed training to make a 1.3bn parameter LLM:…More evidence that distributed training works well …| Import AI
Google DeepMind's Gemini AI officially won gold at the 2025 International Mathematical Olympiad, a feat matched by rival OpenAI, escalating the race for AI reasoning.| WinBuzzer
IMO 2025 just finished. A bunch of AI models gave it a shot this year. Here’s my take on the results.| rishimehta.xyz
A guest post by Walter Dean and Alberto Naibo| siliconreckoner.substack.com
I’m impressed by large language models. So why can't they get the basics of poker right?| www.natesilver.net
Unvibe: A Python Test-Runner that generates correct code| claudio.uk
YouTube link| AXRP - the AI X-risk Research Podcast
Generate code that pass unit-tests| claudio.uk
Could test-time training give AI models this important capability?| www.understandingai.org
We investigate four constraints to scaling AI training: power, chip manufacturing, data, and latency. We predict 2e29 FLOP runs will be feasible by 2030.| Epoch AI
Here I’ll try to explain the coolest ideas in each of AlphaProof’s IMO 2024 solutions. AlphaProof produces proofs in Lean, and each Lean proof is composed of a series of tactics. So I’ll pick out the tactics that correspond to these ideas in the proofs for problems 1, 2 and 6 (the three problems that AlphaProof solved). AlphaProof has developed its own proving style, so figuring out what it’s doing can involve some detective work.| rishimehta.xyz
Industrial Strength Data Science and AI| rssdsaisection.substack.com
What is our responsibility to machines that may become moral patients?| importai.substack.com
Plus: Will billionaires live forever; a police robot dog jamming wireless networks; Alphabet to invest $5B into Waymo; warnings about “model collapse”; a new partnership for AI security; and more!| www.humanityredefined.com