Amodei: Reasoning models (o1-style) trained on synthetic data obviate near-term data scarcity concerns; extended to wider task domains beyond math/coding in coming months

“In the last 6 months there have been some innovations actually not developed by us uh you know first came from uh uh open AI actually um uh but others that that we have made um that uh obiate the need for uh as much data as we need before these are the so-called reasoning models... Um, so far that's mostly applied to tasks like math and computer programming. Um, but my view without being too specific is that it's not going to be terribly difficult to extend that kind of thinking to a much wider range of tasks.”