Agree with your sentiment, though the article explicitly mentions precision/recall, suggesting at least some level of tuning. Query understanding via structured

Agree with your sentiment, though the article explicitly mentions precision/recall, suggesting at least some level of tuning. Query understanding via structured attributes is SOTA and used at top companies. Rewriting the query as a method is weird, and yeah I'm not so convinced. One reoccuring problem - the hacker ethos doesn't scale with AI products. "Mess around until it works" is ok to prototype. This is effectively using the dev's intuition on the 10 examples they look at as the offline eval function. But many (most?) new-wave AI products don't have consistent offline metrics they optimize for. I think this quickly stops working when you've absorbed the obvious gains.