Csaba Szepesvári: ICLR reviewer bidding has become all-negative, no shades of grey
CsabaSzepesvari · x · 2026-09-26
Reinforcement learning researcher Csaba Szepesvári complains that ICLR bidding has become completely negative: reviewers can only choose between "not willing" and "not not willing", with no grades or shades of grey.
He argues the design effectively forces reviewers to express rejection after rejection, stripping out any signal of genuine interest, and questions the sanity of the current bidding system. The post touches on broader concerns about peer-review workflow design at top AI conferences.
Related event: RL Researcher Criticizes ICLR Review Bidding for Negative-Only Options(2 posts)→
More from Research
- RL isn't just sharpening existing skills — pre-LLM RL trained from random init — cephaloform · 2026-09-26
- Steelmanning OPSD: train only on tokens after verbatim rule reminders — willcb · 2026-09-26
- Researchers warn Astra reviews obsess over details, miss the big picture — aran_nayebi · 2026-09-26
- 26 hours, zero prompts: an architecture for self-verifying autonomous AI agents — epicskyes · 2026-09-26
- Self-Play Game AI Trained From Scratch Is 'The Game', Says Researcher; 10x Compute Would Make It Even Better — syhw · 2026-09-26
- If AI solves a 200-year-old math problem today, is that better than waiting? — nathanbenaich · 2026-09-26