I have some quick thoughts that about chat bots built on large language models. TLDR; They produce words in a reasonable pattern matched process, they do not understand or form any value or levels of truth to it and that is the stumbling block that is hitting people now and will continue to hit them with adoption and use.
- Computers remember what you tell them exactly unlike humans (our memories are fuzzy at best and full of shortcuts as memory is malleable and squishy inside our meat muscle called a brain)
- Humans put information in computer that they want preserved and it was assumed to be vetted and accurate as that has been the process we have become accustomed to since their invention (think bills, property records, etc)
- But now AI chat bots create random text and people still operate on the trust of points 1 & 2 so everything is going to be fucked until a lot people realize this until you start distrusting AI chat bots exactly like irrational, uninformed, and liar like humans things are going to be bad for people they are humans that are not like that but to a degree we all are and can't remember things clearly like computers (our dumb fuzzy memory)
- As far as replacing human programmers, I don't think so. If you use the tools to speed up your work it will be better like how people in Star Trek talk to computers to speed stuff up. Starfleet Engineers built a holographic emergency doctor with this approach and a few display pads