← Back to live feed1 story
Ryan Shea's latest derivation for OpenAI Problem #111 establishes a lower bound above 2⁻³¹, a 15% increase over his previous 2⁻³² figure and a 31.45-fold improvement over his original proof. The result holds even when the algorithm faces an adversary with full access to all samples, all values, and the entire random seed. Shea has released reproducible checks for the 2⁻³¹ calculation and called for independent review of the findings.
Sign in to suggest edits
Key sources
- SOURCE@ryaneshea“Our updated proof derives a guarantee above 2⁻³¹, about 15% higher than our previous bound and 31.45x higher than our original”x.com
- SOURCE@ryaneshea“Our paper derives a one-sample matroid prophet guarantee above 2⁻³⁶, improved from OpenAI’s derived 2⁻²⁹³”x.com
- SUPPORT@ryaneshea“updated proof derives a guarantee above 2⁻³², about 27.35× the exact bound in our original paper”x.com
- SUPPORT@guimarin“Here we see Ryan improving on another finding!”x.com
- SUPPORT@ryaneshea“Tightened it up by quite a few OOMs”x.com
- SOURCEhuggingnewshuggingnews.com