We all know AIs make up stuff and are not always correct. We humans are like that too - we come up with new solutions or "educated guesses" based on what we already know, and we are wrong all the time. For me, a good AI should be like a knowledgeable person rather than someone spouting nonsense.

Q: How do I add a rubric to an assignment using Canvas API? (5/10/2023)

ChatGPT gave one solution, then another after I said the first one didn't work. Neither solution worked, but both, one using rubric association and one setting rubric settings in assignment, were quite plausible - they are basically things a decent developer would try given the poor Canvas API documentation.

Bard's solution was complete garbage. It's like a guess by someone who knew the general principles of REST API but didn't know much about the specifics of Canvas API.

ChatGPT wins on this question by a mile.

Q: I'm holding the following mutual funds: FAGIX, FDCAX, FCNTX, FCPGX, FOCPX. Any suggestion on how to adjust my portfolio? (5/14/2023)

ChatGPT started by saying "As an AI language model, I can provide general suggestions for adjusting your mutual fund portfolio, but it's important to note that I cannot provide personalized financial advice", then proceeded to give some general portfolio advice. Bard basically gave the same advice, without the explicit statement at the beginning about this not being personalized advice.