teradimich comments on o1 is a bad idea

[teradimich]

Doesn’t that seem like a reason to be optimistic about reasoning models?