velox_neb

> the naive logic that because LLM execution speeds or benchmarks marginally improved over the last 24 months, (...)

"Marginally improved"? Are you really going to sit there tell me that an appropriate way to sum up the difference between the AI we had access to in Sep 2024 and the AI we have access to now, is "the benchmarks marginally improved"?

> you should not run code you don't trust, period.

This reduces to "you should not use a computer, period." No-one can audit all the code they run daily.

I bet some guy just told Claude Code to archive all of LWN for him on a whim.

My crackpot theory is that what stops newcomers is how unergonomic "(" and ")" are to type on typical keyboards. If mainstream lisp dialects used square brackets instead, we'd all be programming in it!

One does not "simply" use Emacs.

I do, for over a decade. Hope it doesn't die. Can't fathom how people live with these tiny qwerty keyboards on a touchscreen.

Still, most people who have a crypto wallet on their desktop computer don't get hacked. I always found this interesting because it puts an upper bound to how many machines out there are compromised (at least by an agency that's motivated only by money).

Reading these wikis makes me feel we need to invent some visual convention to indicate AI-generated text. Like a particular color or font. This would make it so people don't feel cheated after they realize they just spent several minutes trying to make sense of something churned out by an LLM. (I mean this as a voluntary design enhancement for sites that want to be nice, of course people can always cheat.)

You sound like ChatGPT.

> Microsoft should just buy Everything off Voidtools

Don't give them ideas, they'd find some way to ruin it.

Errata: In the table under "How many bytes are in UTF-8?", bottom row, "10000" should be "100000".

I get "The model: `gpt-4-0314` does not exist".