Stanford professor warns AI-polished grant proposals are getting hard to spot in peer review
anshulkundaje · x · 2026-09-26
Stanford professor Anshul Kundaje argues AI is subtly corroding grant proposals and peer review: if a reviewer is even slightly out of domain, a proposal fully written or polished by AI becomes very hard to pick apart. What bugs him most is the models' confidence — statements detailed and vague enough to sound fantastic unless probed. Push deeper with critical questions and the rationale falls apart quickly, even with "Ultra" thinking enabled.
More from AGI Musings
- Why AI Won't Fully Replace Radiologists: Three Structural Barriers, Explained — gregmushen · 2026-09-26
- Dev shows agents now run his entire self-improving eval loop, echoing Anthropic's recursive self-improvement warning — alexcovo_eth · 2026-09-26
- Why AI Struggles to Replace Radiologists: Long Tail, Sample Size, and Hinton's Miss — gregmushen · 2026-09-26
- Calling fast models "System 1" misappropriates Kahneman, critic argues — emax · 2026-09-26
- Tech firms reportedly seeing meaningful revenue from bring-your-own-agent access — thedealdirector · 2026-09-26
- In the AI era software is treated as disposable, yet we stare at screens half our lives — jeff_weinstein · 2026-09-26