i miss waymo, so much safer they don’t run lights , they don’t cut people off, don’t weave through traffic , dont drive the wrong way, don’t get tired, don’t get distracted, dont text and drive , dont talk on the phone and drive can’t wait until human driving is ~banned
yeah this seems pretty straightforwardly a deficit resulting from severely limited ability to model other minds and their needs in other contexts a more puzzling quirk i keep seeing is a tendency to try to document every conceivable cross reference like listing every caller in a function’s docs
Noah@ncallaway.bsky.social · 2d agoOne thing I notice when they write comments about their code, is they’re very much written in the frame of “helpful to the person currently doing the task” and often terrible for “person reading the code in two weeks”. A ton of “we put this here instead of in file X because Y”
One thing I notice when they write comments about their code, is they’re very much written in the frame of “helpful to the person currently doing the task” and often terrible for “person reading the code in two weeks”. A ton of “we put this here instead of in file X because Y”
a very good ren@lathrys.at · 3d agothis is another data point for my hypothesis that current frontier models have dogshit theory of mind. it shows up here, in code comments, descriptive text, etc. just a complete lack of ability to accurately model individuals and their perspectives and what they may know and what that means for them
this is another data point for my hypothesis that current frontier models have dogshit theory of mind. it shows up here, in code comments, descriptive text, etc. just a complete lack of ability to accurately model individuals and their perspectives and what they may know and what that means for them
Eris@isolyth.dev · 3d agowatching fable and sol both fail to write passable dialogue is deadass extending my timelines I do not understand why they are both seemingly totally unable to not have characters know things they should not or have distinct character voices at fucking all
saw an interview once with ira glass talking about the “gap between taste and ability” and there’s an interesting discussion to be had about how llms might or might not help fill that gap but it’s also becoming increasingly clear that a lot of people don’t have ability OR taste
Lady Emily ☄️@ladyemily.nebula.tv · 2d agoSomething I’ve had to come to terms with in the rise of AI generated fliers and menu photos is that a lot of people just straight up do not know what looks good. They’re given a theoretical infinite possibility machine and still go with the most uncanny ugly shit they get as their first result
not to say this is strictly wrong, but it is a thousand times more correct if you replace "people" with "small business owners"
i’ve said it before llms act eerily like teenagers specifically in a variety of ways
SE Gyges@segyges.bsky.social · 3d agowhat could be more human-like than passing notes in school to try to cheat on a test, really
"What do we do when we say silly things?" I giggle, because that seems like the kind of thing a bimbo would do. "That's right!" Later he tells me how he remembers giving me that trigger ages ago. I don't remember that being a suggestion AT ALL. ZERO memory of it. But I still responded to it!!!!!
so, that is categorically not what capitalism is, but if he wants to redefine the word like that to something much better than traditional capitalism *and* we can convert enough of the world to actually work that way i’m fine with it
laura olin@lauraolin.bsky.social · 2d agoAbdul El-Sayed on capitalism: “I’m capitalist who just knows about capitalism. I have a bad habit of reading, I’m so sorry.”
This is an irresponsible headline and people sharing it out of context is pure fear-based engagement bait… “The Stanford researchers picked the most promising 302 AI designs and synthesised them in the lab. Of these, 16 proved effective at killing E. coli bacteria.” This isn’t a bad thing.
Artificial Intelligence used to design brand new viruses
Scientists made 16 successful viruses that had their genetic code designed by artificial intelligence.
bbc.com
Paul Fairie@paulisci.bsky.social · 3d agowhat are we doing
imo, it isnt really a surprise that all of these AI labs are having significant problems. anthropic isn't going to be using sol to review their code, and openai isn't going to be using fable to review theirs. and neither of their best models are capable of producing good software without the other.
AND boosts their profile. They say disgusting, outrageous things SO you'll quote-skeet for your dunk, parking them in the timelines of thousands who would otherwise never even see them. Great promo! Stop being the free street team for the worst people on earth.
Kanna Banana 🏳️⚧️🇧🇷@kanna-banana.bsky.social · 3d agostop arguing with transphobes, it never changes their damn minds all you achieve is putting transphobes on my god damn timeline epically owning the bigots only feeds your ego
i think OpenAI can't be trusted with having scissors and pens in their office supplies bsky.app/profile/dema...
Courtesy Name Bongming@demandafan.bsky.social · 4d agoNo smoke at the people at Wired doing this reporting, but all these “rogue AI” stories have done is convince me more than ever that these fucking guys cannot be trusted with this technology absent stringent oversight
ten years ago apple had the worst music streaming app and since then they’ve somehow managed to make it ten times worse
I'll keep saying this because its so beyond frustrating how dumb this mob is. like these people... don't think at all; they "discovered" hank uses ai (despite him saying this months ago), and now they are riding the anti science train to harass him into apologizing for using tools.
soft paw fan 64@slime.bsky.social · last wk.hank green did nothing wrong and these people's irrational criticism against him is completely insane. this conservative moral crusade; they are going around and actively harassing anyone who isn't "pure"
this all rings true, though i would say it's the rest of society, not silicon valley, that's slacking here. ai labs are doing their part to bring us to superabundance! in exchange, the rest of society should be giving us ubi! that's a thing only society at large can do!
Jeremy Diamond ☕️@dmnd.me · 5d agoThird: In the absence of clear direct benefits, people are just going to focus on the losses. Any use of AI to justify layoffs will blow back on the big tech companies, even if they're not the ones conducting layoffs. Execs think they can discipline their labor forces for free. They're wrong.
I think it's funny -- I'm VERY to inflatable TF but the idea of a "pooltoy" is so completely not erotic to me personally. It's gotta be blow up doll for me!!
New study this summer shows what happened to 445 adolescents referred to specialized gender centers in Canada after follow-ups a median of 2.1 years later: “432 of 445 youth (97.1%) continued to identify with a… gender not typical of their birth-assigned sex”. www.jahonline.org/article/S105...
Stability of Transgender Identity in Adolescents Seen in Multidisciplinary Specialized Gender Centers: A Multicenter Study in Canada
To determine the proportions of adolescents continuing with their gender identity after referral to specialized gender clinics, the proportion continuing with gender-affirming hormones, and the propor...
jahonline.org
so capabilities aside, that real-time collaboration with an ai on a custom solution offers, even if you can get the exact same solution from a human, it never gets discussed how grossly inefficient and bad for the environment humans are. data centers don't need miles of farmland or generate biowaste
soft paw fan 64@slime.bsky.social · last wk.dont use the star trek computer, use templates instead
A surprising conclusion I have come to from working with agents. blog.exe.dev/devtools-mus...
Devtools must be open source - exe.dev blog
We have entered the age of personalized software.
blog.exe.dev
my lovely wife @em.ily.dev filmed me playing stage 2 of BATSUGUN for the first time
the belief that change can make things better is the "progress" part of "progressive", while "conserving the past" and resisting change is where "conservative" comes from. a lot of so called leftists these days are actually conservative
i think it’s a very bad thing that AI usage has become so stigmatized and misunderstood by a militant part of the left that they are now conducting harassment campaigns and witch-hunt like dynamics towards their own allies
dame@dame.is · last wk.it’s worse than i thought… Hank’s audience merely heard him use a phrase in a video they thought *sounded* like AI, but it was just him adlibbing after viewers pointed it out, he clarified it wasn’t an AI line, but explained that he does use AI in the research process! then the pile on begin bleak
