I tend to agree with you. It's impressive how much ChatGPT gets right, but even as much as it gets right ultimately it's clear when you're reading its output.
I also find it's sort of stubborn such that if it's messing something up it's really hard to get it to stop.
I also find it's sort of stubborn such that if it's messing something up it's really hard to get it to stop.
- replies
- 2
- announces
- 0
- likes
- 0
I've managed to catch it in really silly errors. One time I asked it for a libertarian argument against deregulation of fintech, and it immediately got confused and gave me a libertarian argument for the deregulation of fintech.
(A libertarian argument against the deregulation of fintech would be that because the banking sector is so heavily regulated and so heavily subsidized and so heavily insured by the taxpayer that deregulating fintech without deregulating the rest of finance including removing insurance and government guarantees presents moral hazard that potentially could cause government to increase in scope as the deregulated fintech industries rely more on government insurance and assurances than were originally envisioned by the people who created those programs with only traditional banking in mind.)
(A libertarian argument against the deregulation of fintech would be that because the banking sector is so heavily regulated and so heavily subsidized and so heavily insured by the taxpayer that deregulating fintech without deregulating the rest of finance including removing insurance and government guarantees presents moral hazard that potentially could cause government to increase in scope as the deregulated fintech industries rely more on government insurance and assurances than were originally envisioned by the people who created those programs with only traditional banking in mind.)