I keep continually learning the same lesson: unless your question is simple / superficial / common, the only reliable GPT model is the pro one. I've gotten straight up WRONG answers from normal "oh let me decide which model is correct" mode, switched to pro and.. bam, now it's correct. Depends on the question, of course.

0

If you have a fediverse account, you can quote this note from your own instance. Search https://infosec.exchange/users/codinghorror/statuses/115358739234504635 on your instance and quote it. (Note that quoting is not supported in Mastodon.)