In seriousness,, @suetanvil, it's important to note this. LLM-backed generative AI ballyhoo neglects the fact that Eliza was passing Turing tests (at least, on occasion) before most of us were born.

We're not in the midst of some AI revolution that changes the face of computing. The subfield reached a plateau recently — which has limited uses — except as a capitalist fad, that is — those uses are boundless indeed.

&, @jmax, I can't keep playing M-x doctor b/c my compiles are faster these days!

@suetanvil, fortunately, it would be *both* a violation of GPLv3 and (likely) OpenAI's proprietary license if you hooked up Emacs to Chatgpt's API.

I'm sure M-x doctor would conclude:

“I think your wasteful overuse of electricity & disk space to implement something that is barely better than my 1,625 lines of Elisp has something to do with your problems.”

Cc: @jmax @cwebber

@bkuhn @suetanvil @jmax I'm not sure it is a violation of GPLv3 to hook up Emacs to ChatGPT's API or not, if ChatGPT's service is some REST endpoint (admittedly, I've never looked into whether it is)

If that were true, browsing the web with M-x eww would also get me into a lot of trouble I'd think

Regardless, I don't advise doing it anyway :P

1+ more replies (not shown)
@bkuhn @jmax @cwebber

I'd think that Emacs code which made http calls to ChatGPT's web interface would fall under Freedom Zero, though ChatGPTs TOS is another thing and not something I can really be bothered with.

There *are* models that will run on a desktop PC, so it's not impossible to DIY that end of things either.

Not that there's any value to it, but I kind of like the symmetry of stupid involved in something like this.

@suetanvil, fortunately, it would be *both* a violation of GPLv3 and (likely) OpenAI's proprietary license if you hooked up Emacs to Chatgpt's API.

I'm sure M-x doctor would conclude:

“I think your wasteful overuse of electricity & disk space to implement something that is barely better than my 1,625 lines of Elisp has something to do with your problems.”

Cc: @jmax @cwebber

In seriousness,, @suetanvil, it's important to note this. LLM-backed generative AI ballyhoo neglects the fact that Eliza was passing Turing tests (at least, on occasion) before most of us were born.

We're not in the midst of some AI revolution that changes the face of computing. The subfield reached a plateau recently — which has limited uses — except as a capitalist fad, that is — those uses are boundless indeed.

&, @jmax, I can't keep playing M-x doctor b/c my compiles are faster these days!