Experiment
Topic archive • 3 matches
AI-generated: summaries written by AI from the linked sources. How we use AI
2026-11-14
Technology
A benchmark for wet-lab experiment prediction, described as the strictest of its kind, was released by more than ten top teams including Stanford, Harvard and Cornell. The results indicate that scaling inference and quintupling compute did not improve performance.
AI Research • QbitAI 量子位
Permalink
2026-11-15
Technology
UniPat AI's SPU benchmark asks AI agents to rank the outcomes of 4,207 real wet-lab experiments, and none of the 13 configurations tested scored above 0.5. The best configuration averaged 0.486 out of 1, highlighting how far agents remain from reliably predicting experimental results.
Research • Pandaily
Permalink
2026-11-02
Tips
Prompt Tips
Willison used an Apollo Global Management chart based on U.S. Census Bureau BTOS survey numbers to demonstrate GPT-5 data analysis and to experiment with Claude's Code Interpreter, then re-ran the numbers to see what changed.
Permalink