I already calmed down because it’s quite obvious that OpenAI is engaging in textbook, bait-and-switch startup tactics. GPT-4 performance has noticeably taken a nosedive since its initial release and most recently degraded further in advance of the iOS app release.
Generating worse code in half the time is a service degradation, IMO. I mostly use it for monotonous scripts and one off functions, where saying “write this thing” while I click off the tab and do something else for a minute is a not in anyway a problem. I could tolerate it being 2-4X as slow, because now I have to spend an equivalent amount of time as that correcting errors it didn’t make a month ago.
The worst is changing class names or hallucinating property names. I wasn’t sure if it was just my expectation or if indeed it has gotten worse so it’s good to see other people having the experience.
I have a theory that they made it worse on purpose, not to save money but instead to really train it’s reasoning and arguing skills because I spend so much time ‘fighting with a computer’ now.
