LRM reasoning breaks down down past an unfamiliarity threshold, not "complexity"https://twitter.com/fchollet/status/1933937096286470623 by k1m • 1 year ago 5 0 1 year agoTWtwitter.com