Hi Florian, thank you for the more detailed feedback. Perhaps I am a bit naive about ChatGPT, but there could be future potential here based on what I see today, despite its shortcomings.

WRT LSP, keep in mind that I asked one single question to produce examples for each of the 5 principles. I found that it gave more detailed information when asking about a specific principle separately. And if you ask multiple times it may give different examples and explanations. This is one of the caveats I’ve mentioned WRT learning from it.

I have seen some screenshots of some crazy inaccurate things that ChatGPT can answer (like an abacus is faster than a cpu). I acknowledge the data set is limited to what it’s fed with, and can be subject to bias. However, in my short experience, if you ask it questions both for and against reasonably accepted concepts, it will try to give you both sides of the argument. Though it will side with reason (and with even more “humanity” than humans do), from what I’ve seen. I guess the problem with that is who gets to decide what reasonable and ethical is?