@ph_singer sometimes I want to turn a session in a skill because there so much good findings in there and I am 99% sure the situation for this skill will it happen again, but if it will Claude will just one shot the solution
@trq212 I am curious if you have any logic around TTL of caches? Are all calls cached with default 5m TTL or if you detect that the plan mode was triggered and the generated plan is more than e.g 1000 tokens it might take user more than 5 minutes to read through you would use 1h
@rodoss В 2012 вдохновившись его просмотром решил сходить на экскурсию в студию NBC на 30 Rock и когда заговорил с экскурсоводом про этот сериал никто больше из группы его не знал, но все знали 30 Rock. Даже обидно немного стало за него
Our open machine learning course starts in a week https://t.co/YC5Y0QHHfx (for russians also https://t.co/RoVbfQpGGs). 10 articles, assignments and several @kaggle competitions. @odsai_en