The question of what AI companies would do when they ran out of training data seems to have been answered. As expected, but with a twist, they’ll train new models on the results produced by hyper-accurate models that come before it.

The question of what AI companies would do when they ran out of training data seems to have been answered. As expected, but with a twist, they’ll train new models on the results produced by hyper-accurate models that come before it. https://www.lesswrong.com/posts/8oX4FTRa8MJodArhj/the-information-openai-shows-strawberry-to-feds-races-to