A korrentour readingWhat is a korrent?
Post-training is what turns a model into a product, but it can only do that on top of a pre-trained model large enough to already have common sense.
Drawn from what Aravind Srinivas said
Subjectscaling laws
scaling laws Whether more compute keeps buying more capability, and where the curve now bends -- pre-training, post-training, test-time.
A private bookmark. Not a position, and never counted.