Pinned
we started a company!!
so, we’re tackling continual learning: what’s the learning algorithm to take arbitrary data — documents, conversations, the models’ own experience — and make better models? how do we scale compute in the same way we’ve already seen with pre-training and
