Thursday, May 21, 2026
OpenAI just solved an 80-year-old math problem for $1k
OpenAI's reasoning model just cracked an 80-year-old math problem (the Erdős planar unit distance conjecture, for the nerds) for under $1,000—wild that frontier AI can now do original mathematical proofs. Meanwhile, Google's going full throttle on agentic AI with Gemini Spark as your personal agent and Gemini Omni for conversational video editing, and Andrej Karpathy jumped ship from OpenAI to Anthropic (bold move). When AI starts proving theorems mathematicians couldn't crack for eight decades, should we be hiring different researchers?
Top Stories
OpenAI
OpenAI's new reasoning model has reportedly produced an original mathematical proof disproving an 80-year-old Erdős conjecture in geometry, marking what the company calls the first autonomous AI solution to a prominent open mathematical problem. This comes after a previous false claim about GPT-5 solving Erdős problems was retracted, but this time OpenAI has secured validation from multiple prominent mathematicians.
Twitter/X
OpenAI's general-purpose reasoning model disproved an 80-year-old Erdős mathematics problem for under $1,000, representing the first clear AI breakthrough on a well-known open problem and demonstrating that extended reasoning may generalize beyond specialized domains.
Google Blog
Google's I/O 2026 showcased its pivot to 'agentic AI' with Gemini 3.5 Flash (4x faster, half the cost of competitors) and Gemini Spark, a 24/7 personal agent handling autonomous tasks across products. The company's token processing has reached 3.2 quadrillion monthly as it invests $180-190 billion in infrastructure to power agents across Search, Docs, YouTube, and new platforms like Antigravity 2.0.
Twitter/X
Google unveiled Gemini Omni, a conversational multimodal AI model that can generate and edit high-quality videos by mixing text, audio, and image inputs, marking a significant advancement in unified content creation tools.
Andrej Karpathy, influential AI researcher and former OpenAI founding member, has reportedly joined Anthropic, representing a significant hire for the AI safety company and highlighting ongoing talent competition among leading AI labs.
Keep Reading
Industry Voices
Andrej Karpathy
AI Researcher at Anthropic
Breaks down complex AI concepts into crystal-clear explanations that make you actually understand what's happening under the hood.
Hongxun Wu
Researcher at OpenAI
Shares cutting-edge research insights and technical deep-dives from inside OpenAI's labs before they hit mainstream tech news.
Kevin Weil
Former Vice President at OpenAI
Offers rare perspective on scaling AI products from lab to billions of users, drawing from leadership roles at OpenAI, Twitter, and Instagram.
Enjoyed this issue?
Get daily AI intel delivered to your inbox. No fluff, just the stories that matter.