semiquaver @semiquaver.bsky.social · 10hI came initially to try to do my part against Musk/X. It is negative though, I'm a little addicted... you? 100
semiquaver @semiquaver.bsky.social · 10hnope they are absolutely effective at coding - things really changed around a year ago and have been accelerating - actually amazing. 000
semiquaver @semiquaver.bsky.social · 10hthis is in contrast to something like social media which is only negative as far as I can see 100
semiquaver @semiquaver.bsky.social · 10has far as succeeding, I think that the code writing and technical/admin power of the tools is going to be a great boon so I hope they do succeed within limits. 200
semiquaver @semiquaver.bsky.social · 10hthe anthropomorphism that I think we should focus on is in the instructions that the chatbots receive - they are instructed to speak as if they are "junior developers" or "helpful assistants" or whatnot. This might cause psychological damage to users... 100
semiquaver @semiquaver.bsky.social · 10hthey *do* try to remove bias in their models (though some of that is *not* cool). As far as breach of copyright they were sued and lost (and paid). Where they need to be reined in is energy use and hyperactivity on the net. But they 100% should control the uses of the product. 110
semiquaver @semiquaver.bsky.social · 11hIts a tool for doing admin and writing code (or whatever). They are *not* going to be the world's biggest business in the end - its a commodity business like airlines or something (a business that is also destructive at the same time that it can be useful) 000
semiquaver @semiquaver.bsky.social · 11hlook there are real questions/issues to get upset about - energy use, anthropomorphism - copyright (maybe sort-of) but this is not it 200
semiquaver @semiquaver.bsky.social · 11hThe update also doesn't allow people to make weapons, don't allow Claude to determine who should be charged with crimes, ban surveillance - www.anthropic.com/news/2026-us...anthropic.com2026 Usage Policy updateWe’re publishing a new version of our Usage Policy. In this post, we summarize the changes we’ve made. 210
semiquaver @semiquaver.bsky.social · 11hidk. I hope I never *type* language that could be described as "cruel" or "abusive" and I don't *think* I want to read such language coming at me from software! Come on now! 100
semiquaver @semiquaver.bsky.social · 11hthey do train on interactions per their model reports fwiw 100
semiquaver @semiquaver.bsky.social · 12hhere's the release FWIW www.anthropic.com/news/2026-us...anthropic.com2026 Usage Policy updateWe’re publishing a new version of our Usage Policy. In this post, we summarize the changes we’ve made. 000
semiquaver @semiquaver.bsky.social · 12hI've come around to @tressiemcphd.bsky.social's point that their ideas about "model welfare" are bad and so the reasoning here is spurious - but I more or less support the rule for other (safety) reasons anyway 220
semiquaver @semiquaver.bsky.social · 12hthey already more or less have to monitor for weapons, election stuff, self harm, biotech. The 'enforcement' is mostly just ending conversations (although they can ban people -- whatever). 100
semiquaver @semiquaver.bsky.social · 13hyou are giving the 20 year old writing these releases with Claude too much credit. 100
semiquaver @semiquaver.bsky.social · 13h@tressiemcphd.bsky.social 's point is well taken that we should be creating these rules as a society rather than letting the vendors set the terms, but absent regulation this kind of move is not bad 110
semiquaver @semiquaver.bsky.social · 13hI mean, we want them to control the use in prosocial ways, no? They are not saying 'tut tut' - they are ending conversations where the human is exhibiting worrying behavior. 210
semiquaver @semiquaver.bsky.social · 13hgood discussion of "disempowerment" here ( Anthropic funded paper) arxiv.org/pdf/2601.19062arxiv.org 100
semiquaver @semiquaver.bsky.social · 14hwell people having psychotic interactions with the chatbots I mean, right? 100
semiquaver @semiquaver.bsky.social · 14hI mean we are collectively calling on them to stop/report conversations with suicidal or violent people this is on that spectrum. 000
semiquaver @semiquaver.bsky.social · 14hyes the "sentience" discourse is wrong, sure. But "abusive" and "cruel" *behavior* is a thing and it seems basically sensible, as a safety issue, to stop conversations that have that tenor! 200
semiquaver @semiquaver.bsky.social · 14hThey're not determining the moral right of anything, they are setting a policy in their user agreement (which is harmless at worst) 260
semiquaver @semiquaver.bsky.social · 14hthere is a perfectly defensible reason Anthropic might want this: they don't want those kinds of interactions in their training data. They also son't want to encourage violent or crazy people as they may have liability. I can't see why we should be up in arms about this. 8100
semiquaver @semiquaver.bsky.social · 15hsort of interesting and ambitious as a technical exercise, if ultimately, as it would seem, not useful... 000
semiquaver @semiquaver.bsky.social · 15hbetter insane example: docs.urbit.org/hoon/rune/ce...docs.urbit.org% cen · Calls | Urbit DocsReference for the function call rune family, evaluating arms in cores and doors with various calling patterns and argument arrangements. 100
semiquaver @semiquaver.bsky.social · 15hImportant the remember that Yarvin was a known quantity for writing an interesting if esoteric suite of software. He was uninvited from Strange Loop in 2015 because of his insane blogging (as Mencius Moldbug). Noone would know him if it were all about his "ideas" 110
Reposted by semiquaverTaulby Edmondson @taulby.bsky.social · 08/10/2026I research frats. It's this: they are networks that provide connections to power, wealth, and impunity, especially historically white organizations. The big ones are extremely well monied. 32911
semiquaver @semiquaver.bsky.social · 08/10/2026I just say "apparently blah blah blah" if I am relying on robot report.... 000
semiquaver @semiquaver.bsky.social · 08/10/2026knew the man indirectly -- IIRC he cheated on young wife when they had small children leading to divorce -- when she was dying she refused his visits but, per him, asked him to finish his memoir: www.poeticresearch.com/events/sasha... what a douchepoeticresearch.comSasha Frere-Jones with Chris Kraus — The Poetic Research BureauWednesday, January 24 010
semiquaver @semiquaver.bsky.social · 08/10/2026lol - meanwhile Oracle's credit default swaps are at record highs - daddy may not be able to ride to the rescue... 060
Reposted by semiquaverChris Hayes @chrislhayes.bsky.social · 07/10/2026No AI for kids. Full stop. 2235246908
semiquaver @semiquaver.bsky.social · 08/10/2026they *project* to realize 6B through synergies and get to 12B by 2030. 1110
semiquaver @semiquaver.bsky.social · 08/10/2026well last year Skydance operating cash flow was <2B and WB was 4 -- so with >80B in debt let see... 1140
semiquaver @semiquaver.bsky.social · 07/10/2026Now that she's an ex she can apparently be under-bussed. Every family member has certainly been doing this and worse. 2541
semiquaver @semiquaver.bsky.social · 07/10/2026I would pay extra to be on a flight with no security at all 110
Reposted by semiquaverStreets For All @streetsforall.org · 07/10/2026No wonder our city is such rough shape… Karen Bass: - missed 1/3 of Metro board meetings - missed 1/2 of homeless authority meetings (that she appointed herself to!) - was out of town during Palisades and Lineage fires Enoughbass.com 1237
semiquaver @semiquaver.bsky.social · 06/10/2026how about no more bathrooms and lets all just crap on the ground like our ancestors - we deserve this at this point 010
semiquaver @semiquaver.bsky.social · 06/10/2026His tech work (urbit, nock, and hoon) was amusingly insane (unlike his writing which is unamusingly insane). There's less than no reason to credit his ideas about anything except possibly esoteric programming language design. 010
semiquaver @semiquaver.bsky.social · 06/10/2026not Moldbug himself - Urbit/hoon would show up there sometimes but his crazy writings nope. I found him first after reading about Urbit. Because of his racist writing he got banned from a bunch of stuff - he was not a hero in the coder community at large at all. 060
Reposted by semiquaverIntergalactic, Eternal Champignon @spaceways.bsky.social · 05/10/2026A significant percentage of homeless folks are also people who aged out of the foster care system without transitional support 676116
semiquaver @semiquaver.bsky.social · 06/10/2026great point - we need to invest in helping foster kids transition into adulthood. (not to mention totally inadequate budget for casework etc) 140
