Just to be clear: this was said by Bryan Catanzaro, the VP of applied deep learning at Nvidia. It is not generalizable to workers using AI tools on their day-to-day work.
A few teams like his are notorious for requiring a ton of computation as they try out different novel approaches, particularly in the context of training rather than inference.
Bryan will sometimes comment on HN, so we may get lucky.
david-gpu•1h ago
A few teams like his are notorious for requiring a ton of computation as they try out different novel approaches, particularly in the context of training rather than inference.
Bryan will sometimes comment on HN, so we may get lucky.