Hacker News new | ask | show | jobs
by mmmore 527 days ago
The comment was likely that there's no explicit search. In o1, the model has learned how to search using its context. Presumably they do this by RLing over long reasoning strings/internal monologues.