antifuchs

joined 2 years ago
[–] antifuchs@awful.systems 10 points 1 day ago

Here’s a substack post (sorry) with a quote I found both neat and pretty funny:

Integrity comes from the Latin "integer," meaning whole or complete. A person with integrity is "whole" in the sense that their words, actions, and values are unified rather than fragmented or contradictory. They understand themselves; they have integrated the warring parts of themselves; and they respect and act on the values that their parts can agree upon.

Rationalists in shambles

[–] antifuchs@awful.systems 8 points 4 days ago

Something intensely satisfying seeing this the same week as id software becoming a union shop.

[–] antifuchs@awful.systems 11 points 2 weeks ago (1 children)

Workers organizing against genai policies in the workplace: http://workersdecide.tech/

Sounds like exactly the thing unions and labor organizing is good for. Glad to see it.

[–] antifuchs@awful.systems 14 points 2 weeks ago (1 children)

They’re doing eugenics ads on the nyc subway https://hexa.club/@phooky/115640559205329476

[–] antifuchs@awful.systems 6 points 2 weeks ago

Yup that’s one of them. The cryptonomicon protagonist no-nut-Novembering all the way to the ww2 treasure is another special fave

[–] antifuchs@awful.systems 12 points 2 weeks ago (5 children)

Lmao imagine reading a Stephenson book and being peeved that it ends

(His sex scenes are far far far worse than his endings, those are a mercy)

[–] antifuchs@awful.systems 3 points 3 weeks ago* (last edited 3 weeks ago)

Ah yes, the classic open source guy stance of “you get to praise me publicly but direct criticism or problems to [over there]”

[–] antifuchs@awful.systems 15 points 3 weeks ago (4 children)

PR for AI slop generated DWARF support for ocaml. Expectation: doesn’t work

Reality: replicates existing support from another project including attribution to the author.

[–] antifuchs@awful.systems 7 points 4 weeks ago

I would love this, lmfao

[–] antifuchs@awful.systems 14 points 4 weeks ago (2 children)

Claims to be against politics in the workplace, takes down rightwing websites anyway

[–] antifuchs@awful.systems 3 points 4 weeks ago (10 children)

This might be one of those “careful what you wish for” scenarios. The neutral outcome might well be that they generate dialog with training data from Scots Wikipedia. Or Scots Grokipedia I guess

[–] antifuchs@awful.systems 7 points 1 month ago

Oh look at that, another report on the economics of ai datacenter buildouts https://publicenterprise.org/report/bubble-or-nothing/

 

Got the pointer to this from Allison Parrish who says it better than I could:

it's a very compelling paper, with a super clever methodology, and (i'm paraphrasing/extrapolating) shows that "alignment" strategies like RLHF only work to ensure that it never seems like a white person is saying something overtly racist, rather than addressing the actual prejudice baked into the model.

 

School student tells AI to put 20 other students’ faces on nude pictures, shares them in chat; it takes months for anyone including the school administrators to act because of some extremely, uh, dubious loophole.

If someone does that in photoshop, it’s a crime; if they do it in AI pretending to be photoshop, it’s somehow not. Gotta love this legal system’s focus on minor technicalities rather than the harm done.

 

They have Nik Suresh (the author) on, as well as Robert Evans. I haven’t listened to it all yet, but it’s fun so far.

 

They invited that guy back. I do have to admit, I admire his inability to read a room.

view more: next ›