Comment by sky2224

19 days ago

I should probably do that more, however, one big issue is I can't always trust that I'm not being led astray with some of the posts on Chegg or the responses from ChatGPT and other AI model interfaces. I will frequently get responses from ChatGPT and/or Chegg that are just flat out wrong. So I could compare my answers against these services (which takes a long time), but I run the risk now of potentially learning the wrong approach when I go and find out that the walkthrough I was following was incorrect.

Which granted, that's just part of the work, but these weird hoops to jump through just shouldn't be there in the first place, and they used to not be there as well. So my gripe is why on Earth did we add them?

Yeah, the new interface sounds ridiculously bad. Sorry you have to go through that.

o1-pro is quite reliable for elementary problems, FWIW. I use it extensively for reading AI research papers.