Whether the 80% horizon is rising fast depends entirely on the fitting method
BasilHalperin · x · 2026-11-09
BasilHalperin wraps up his thread's key point: whether OpenAI's 80%-accuracy time horizon is rising fast depends entirely on the fitting method. Fitting the empirical p80 "naively" shows slow growth (585-day doubling); fitting METR-style against success curves shows doubling every 2 months.
He considers the METR-style fit more comparable but remains unsure what to make of the discrepancy between the two results.
More from Research
- LoRA-tuned Qwen3 skips first output line whenever the caption is under ~4 tokens — prathamdmehta · 2026-11-09
- Ramez: OpenAI's Productivity Gain Per ECI Point Appears to Decline Over Time — ramez · 2026-11-09
- Betting on transformers: incremental improvements will beat radically new LLM architectures — max_paperclips · 2026-11-09
- FragBench benchmark paper shared, details in the paper itself — arthurcolle · 2026-11-09
- Economist Paul Novosad proposes a 4-step AI-assisted peer review workflow — paulnovosad · 2026-11-09
- CoRL 2026 to host 'Do Robots Need World Models?' debate workshop in Austin — YunzhuLiYZ · 2026-11-09