That's not the part that's different. The part where I feel we perhaps differ is rather than being "premium mediocre" I think that openAI is really excellent where the problem space is very broad or is poorly specified. Then we both agree there are better choices where it is narrow and well specified.
