OpenAI researcher posts slides on disproving the exponential time hypothesis
SebastienBubeck · x · 2026-10-29
OpenAI researcher Mark Sellke (reshared by Sebastien Bubeck) has posted slides outlining OpenAI's approach to disproving the exponential time hypothesis. He notes the outline skips some steps, yielding a smaller power of log(n) in the exponent than the fully optimized algorithm — a rare public look at complexity-theory work inside a frontier AI lab.
More from Research
- Stanford colloquium: choosing insight over overload in the age of AI — jonc101x · 2026-10-29
- A 2-neuron +0.75 tweak makes 'evil Qwen': So8res on what steering really reveals — builderjaydub · 2026-10-29
- AdamW batch-size scaling rule: scale LR and weight decay by √k, validated at ICLR — orvieto_antonio · 2026-10-29
- AMARO v2.0: one neural potential for proteins, ligands and complexes, 3.3x more pose recoveries — gdefabritiis · 2026-10-29
- Nonprofit lab releases Agaram 2.5B, the first Tamil model trained from scratch with tool calling — SuB8u · 2026-10-29
- NUS Proposes Wiener-Informed Decoding to Speed Up Pixel-Space Flow Model Training — NationalUniversityofSingapore · 2026-10-29