Researcher slams papers whose stacked changes beat even weak baselines barely
retr0jirachi · x · 2026-10-16
A researcher vents about a common pattern in AI papers: multiple changes to the current paradigm presented as a major breakthrough, only for the baselines to reveal that even plain behavioral cloning would have achieved nearly the same performance. The critique targets weak baselines and stacked ablations that obscure how little the actual contribution is.
More from Research
- Square lattice connective constant remains unsolved — ripe for AI, argues mathematician — tak3sh8 · 2026-10-16
- Quiz: Double the Residual Stream from 4096 to 8192 — How Does Param Count Change? — Sauers_ · 2026-10-16
- Researcher credits GPT for multiple accepted papers, incl. disproving Wegner's 1965 conjecture — ctjlewis · 2026-10-16
- AutoAgent0: VLM-powered agentic runtime lifts long-tail driving success by 28.3pp — yining_hong · 2026-10-16
- Shuffled labels hit AUC 0.59: selection bias, not signal, drives prosody benchmarks — maier_ak · 2026-10-16
- CODESTRUCT reframes codebases as AST action spaces to fix brittle LLM code agents — burkov · 2026-10-16