Accumulating data avoids model collapse, and an untold AI-agents backstory
RylanSchaeffer · x · 2026-10-20
Rylan Schaeffer reflects on "Is Model Collapse Inevitable?": unlike prior work where data is replaced each iteration, the paper shows model collapse is avoided when real and synthetic data accumulate across generations — a key correction when the field feared synthetic-data doom. He also shares an never-publicly-told 2024 backstory about AI agents contributing to novel ML research, a direction he regrets was forced to drop. On his NeurIPS-winning "Emergent Abilities a Mirage" paper, he says results are valid but obvious and already known inside labs.
More from Research
- OpenAI claims proof of the quasi-Riemann hypothesis, and human mathematicians have verified it — akbirthko · 2026-10-20
- Albumin is the body's natural defence against black fungus, Nature study finds — realmeetjames · 2026-10-20
- Sonia Joseph on The Information: Are We Building World Models the Wrong Way? — soniajoseph_ · 2026-10-20
- 125,091 DICOM files from a cancer journey released publicly for deep learning research — typewriters · 2026-10-20
- Preprint maps 280 conserved fibrosis enhancers via CRISPRi CROP-seq across 549,767 cells — anshulkundaje · 2026-10-20
- Jason Wei: AI won't automate radiologists — elastic demand beats displacement — _jasonwei · 2026-10-20