Tokenflation: When "Hi" triggers 33 tool callsquesma.com 6 pointsjakozaur1 month ago1 commentSaveHideCopy link On HNComments−cyanydeez1moi can't believe people are waiting 5 minutes for a hi call on a cloud model. I assume most of that is provisioning; local model from cold start only takes ~30s
Comments
i can't believe people are waiting 5 minutes for a hi call on a cloud model. I assume most of that is provisioning; local model from cold start only takes ~30s