IRC channel logs
2026-08-18.log
back to list of logs
<old>people will probably run local LLM that are opensource on their machine <old>but the econimic model of openai and anthropic is doom imo <loon>bremner: "it doesn't pass all the tests, it isn't how I'd write it and isn't idiomatic, it's slow af, but it's fine :)" <bremner>anyway, I have other channels to rant about LLMs in ;). <old>FYI, I did that already in for a project called the Exatracer, without LLM. It was clean and small. Yet somebody in AMD thought it was a good idea to do a 160 commits PR for some reason <old>and I remember a discussion with the person behind this PR that brag about how many agents they have used to make it. "Reviewed by 6 different agents" or something like this <old>if this is the future of our craft, I will go back to university to become a biologist or something <loon>at that point, it's no longer a craft <civodul>bremner: ouch, from Flatt even; “programs are written for humans to read” they said… <old>apteryx: I thnk we move to guix to host the doc <old>tho I get a 403 error here <loon>is it in the wayback machine? <apteryx>probably, I was just surprised to not find what I thought was still the "canonical" doc URL in web search engines <rlb>I'm assume we probably can't establish redirections for the old links? (Be nice to keep them working for now/ever if we can.) <old>You can also participate on the mailing list <old>All inputs are welcomed <mwette>For the next release of nyacc (4.0), ffi-helper generated code will depend only on guile-cdata (or some other backend), not on nyacc itself. <mwette>nyacc-4.0 will include experimental backends for using scheme-bytestructures or bstructs instead of cdata. <old>Is there anybody that could endorse me on arXiv for cs.SE or cs.OH ? I have an article from my master thesis I would like to publish there