zlacker

>>djtrip+(OP)
Yes, ban it. I've been playing around with ChatGPT and where it starts failing is just where things start becoming interesting. What that means is that it's wikipedia-smart, i.e. it doesn't really tell you anything you can't find out with a minimal Google search. It does however cut the time-to-answer quite a bit, particularly if it's an area of knowledge you're not really that familiar with. But it bottoms out right as things start getting interesting, expertise wise.

Case example: I tried seeing what its limits on chemical knowledge were, starting with simple electron structures of molecules, and it does OK - remarkably, it got the advanced high-school level of methane's electronic structure right. It choked when it came to the molecular orbital picture and while it managed to list the differences between old-school hybrid orbitals and modern molecular orbitals, it couldn't really go into any interesting details about the molecular orbital structure of methane. Searching the web, I notice such details are mostly found in places like figures in research papers, not so much in text.

On the other hand, since I'm a neophyte when it comes to database architecture, it was great at answering what I'm sure any expert would consider basic questions.

Allowing comments sections to be clogged up with ChatGPT output would thus be like going to a restaurant that only served averaged-out mediocre but mostly-acceptable takes on recipes.

>>photoc+K8
The problem with ChatGPT is that it often reads authoritative. But is often just flat out wrong.

I asked it a few questions for which I consider myself a subject matter expert and the answers were laughably wrong.

>>scarfa+Km1
I just asked it to provide a specific example of a security breach caused by a Java serialization flaw and it just made something up, i.e. it found some random security breach (unrelated to anything Java) and found some other random Java serialization bug and claimed that breach was due to that bug. A few minutes on Google revealed they were completely unrelated. Ouch.

>>photoc+Gp1
It seems like if the contextual analogy here was carried through and ChatGPT were to leave its message as a comment on a Java-security thread, the same approach would apply (a few minutes of research on Google) and the provably-incorrect message would either be downvoted or commented on, just like a human comment with the same content would be.