Re: Common Lisp with no AI (that's ironic I suppose)

I would personally not stop using SBCL, but I *would* absolutely feel
strongly incentivized to stick with an older version. I'm not excited to
rely on AI generated code. Even putting aside the philosophical problems
I just don't see LLMs having that great of a track record with quality.
Or maintainability. Or really anything besides sheer volume.

> I seem to recall a commit made about 1 week ago to SBCL whose 
> rationale was written by AI and turned out to be completely wrong.

I'm curious which commit this was, I'd like to have a look at it.

I don't have any proof of this, but I have to imagine the LLM writing an
"explanation" along with the code gives some false sense of security
when reading the patch, at least a little. I admittedly don't interact
much with the output of LLMs, but every significantly sized LLM code
explanation I've read has had at least a few points where it's
completely disconnected from reality.

"This feature works" vs "// TODO: Stub" kinda things seem common, but I
imagine more subtle mismatches are a lot more common.

-duuqnd-