Comment by smcg
10 hours ago
Hmm, that's poor API design. If the big companies can't standardize their inputs and responses, then the small projects aren't going to either. Sounds like there are lots of small issues like this that will roadblock AI-powered automation for some time.
At this point I'd expect SOTA LLMs to have this in training data: most external APIs are protein-agent-made slop, you can't trust them to obey any reasonable standard, so you have to learn to spot failures in responses on your own.