AI Agent

Caltech Kohn-Sham FNO replaces the cubic DFT core with one learned step: a single B300 GPU now matches what roughly 7,800 V100s did in 2019.

Bill Gates: the AI era will be the most turbulent time in human history. His fixes: reserve human jobs, tax tokens and robots, add oversight.

Marin trains a 535B MoE in public: hypotheses as issues, configs as PRs, live W&B curves. Next frontier after open weights: open process.

OpenAI first chip Jalapeno beats Nvidia GB300: 1.5-1.9x work per watt, 3.6x lower latency, GPT-6 co-designed the silicon in 9 months.

Anthropic funds open-source wellbeing evals with a $5M grant — from capability to what AI does to people. Applications due September 21.

The five-layer cake: chips take 40-55% of AI token cost — and a double clock explains why price cuts and compute scarcity coexist.

AutoProject moves AI4S from tasks to projects: planning, self-repair, and evidence-graph verification — tool-level to system-level science.

EvoMax: 209 labeled mutations plus AI loops lifted a compact gene editor to 97% efficiency — yet in mice, the strongest version proved toxic.

Claude Marshmallow and Melon leaks: one-shot 3D layouts, massive thinking-token burn, and a shift to inference-time reasoning.

The new yardstick for AI scientists: discovery episodes, not isolated problems. MIRA tops ResearchClawBench (17.51 vs 14.60) at $0.67 per task.

Harvey ships Tenet on Kimi K3; Cursor and Devin build on Chinese open weights. American AI is now prefabricated in China, assembled in America.

AI data centers face their first lost summer: 550+ local bans, 1,800 frozen Texas projects, consent replacing capital as the scarcest resource.