Tokenless

Mathematical reasoning

Inside OpenAI’s Breakthroughs in Mathematical Reasoning

Inside OpenAI’s Breakthroughs in Mathematical Reasoning

OpenAI mathematicians discuss how AI, particularly models like Astra and GPT-5, are rapidly advancing mathematical capabilities. They delve into recent breakthroughs in sphere packing and the discovery of a non-sofic group, emphasizing that AI's reasoning traces remarkably resemble expert human thought, including strategic decision-making and backtracking, rather than mere brute-force. The discussion also explores the changing role of human mathematicians, AI's emergent 'taste' in problem-solving, and the potential for a new era of accelerated mathematical discovery and understanding.

The Evolution of Computers

The Evolution of Computers

This conversation explores how AI's mathematical progress challenges traditional computing assumptions. It examines the shift from engineering-bound problems to capital-bound problems, impacting startups, incumbents, and venture capital, and considers the unpredictable potential of massively funded AI models.

How a reasoning model cracked an 80-year-old math problem — the OpenAI Podcast Ep. 20

How a reasoning model cracked an 80-year-old math problem — the OpenAI Podcast Ep. 20

OpenAI's reasoning researchers discuss how a general-purpose AI model disproved an 80-year-old conjecture from mathematician Paul Erdős. They detail the journey from initial IMO/IOI breakthroughs to the verification of the proof, highlighting the model's creative application of advanced number theory. The episode explores the profound implications for the future of mathematics, AI-human collaboration, and the broader scientific landscape, offering advice for researchers seeking to leverage AI for groundbreaking discoveries.