x.com/daniel_mac8/status/2083339147637305488
1 correction found
they've solved 10 unsolved maths problems
OpenAI did not announce 10 fully solved math problems. Its August 1 release described ten advances or results on long-open problems, including partial bounds and lower-bound improvements.
Full reasoning
OpenAI’s own August 1, 2026 announcement does not say it solved ten unsolved math problems. It says it is sharing "a selection of ten results to problems" that had been open for at least a decade. The listed items include partial advances such as new upper bounds on sphere-packing density, improved bounds on binary and spherical codes, and new lower bounds for arithmetic circuit complexity. Those are meaningful research results, but they are not all complete problem solutions.
So the post overstates what was announced: the release was about ten advances/results across mathematics and theoretical computer science, not ten outright solved unsolved math problems.
1 source
- Ten advances in mathematics and theoretical computer science | OpenAI
OpenAI says it is sharing “a selection of ten results to problems” including “New upper bounds,” “Exponentially improved bounds,” and “New lower bounds.”