Not really possible to cache. The model takes as input the entire chat (or some version of personalized history i guess). But it s more likely that bing will d

Not really possible to cache. The model takes as input the entire chat (or some version of personalized history i guess). But it s more likely that bing will discriminate between simple "keyword queries" and elaborate quesitons