Debate: correct serial baseline for team-of-N is best-of-1 at N*T budget
generatorman_ai · x · 2026-09-22
Responding to plots comparing team-of-N against best-of-N individual agents, @generatormanai argues the fair serial baseline for team-of-N at token budget T is best-of-1 at NT, not best-of-N at T.
More from Research
- SkillLift cuts agent skill-evolution token cost 40-70% by ranking, not rollouts — dair_ai · 2026-09-22
- FAIR & NYU scaling law paper pushes fits down to 4M-param models — giffmana · 2026-09-22
- Salesforce: Fine-tuning a weak model to copy Gemini drops success 15%; correcting its own failures works — rohanpaul_ai · 2026-09-22
- Jevons effect sparks encoder model revival: 'BERT revenge' over GPT — JFPuget · 2026-09-22
- Interview: first conscious AI may 'leapfrog' to sophisticated consciousness — eschwitz · 2026-09-22
- Human video emerges as robotics pretraining substrate: effective experience = hours × information per hour — rohanpaul_ai · 2026-09-22