Sam Altman

@sama

RT @thekaransinghal: One of the ways scaling test-time compute can benefit people most: have reasoning models think really hard about rare…