Humans didn't make slop historically! It resulted from SEO demanding people including stories or pics in their cooking recipes to get higher listing. There was no market for slop in 1820. AI making more slop is tragic because no one ever wanted it; it was an externality that polluted the commons.
GOOD HEAVENS!
It's as if working one's ass off to *not* be able to afford basic needs while watching our freedoms whither away as they are consolidated to an uber-wealthy oligarchy and people die of preventable disease while AI takes jobs is depressing!
news.gallup.com/poll/708221/...
Alexander Kustov @akoustov.bsky.social • 8h For folks who still believe it's OK to stay on this website, I'm genuinely curious: can you go through these quote posts and tell me with a straight face that this is an acceptable way for adults with PhDs to engage in intellectual discussion in 2026?
This has real “grandma can you pick me up from the rap battle” energy
Good critical analysis of the idea that AI could end humanity by @lizziegibney.bsky.social
Many "warn that raising the alarm unnecessarily could be harmful" by distracting from well-documented risks of AI — such as misinformation and mass surveillance.
www.nature.com/articles/d41...
Remember Chomsky's arguments about language innateness based on its putative un-learnability? Those arguments hinge on a pretty crazy theory of grammar. Here is Ted Gibson's summary of the original claims and a proposed solution in the form of dependency grammar:
tinyurl.com/yzjrpbb8
New pre-print alert! How do children learn magnitude words like "long" and "high", which often denote multiple domains? With @urvi.bsky.social and @drbarner.bsky.social, we find that children start with narrow meanings restricted to the labeled domain, before analogically extending
osf.io/ucxra_v1
The rise in climate pollution from the US is 2025 canceled out the reductions from China, India, Japan, and the EU combined.
interesting paper on what makes a psychological event count as learning by henry schiller, who is not on bluesky. highly recommend it
philpapers.org/rec/SCHRLP-4
AGAINST ALL ODDS, THOSE WHO BELIEVED BLUESKY THAT THE CRASH WAS DUE TO A DDOS
AND THOSE WHO DIDN'T BELIEVE BLUESKY AND WERE SURE IT WAS BECAUSE OF THE AI VIBE CODING
WERE
BOTH
RIGHT
The fact of the matter is that your racist uncle who hates Woke Studies ever since he heard about it on the drive back from the supermarket and the governing board that oversees your research institution are converging, and the results are exactly as monstrous and as stupid as you would have guessed
Some people have misunderstood this.
We literally saw how Claude itself, coded by Claude, turns one API call into a DOS of thousands upon thousands of API calls done in the sloppiest fucking way possible.
This postmortem shows Bluesky vibecoded a function to DOS itself, in a similarly stupid way.
If AI was basically popular and useful but we couldn't monetize it, that'd be one thing. If we all loved the games and community and connections, great. That's not really the issue. It's not broadly useful, it's costly and environmentally damaging, and it destroys jobs without replacing them.
This is true but a bit confused. The Internet was popular and massive but hard to monetize in 1999. The dot com bubble was about monetizing it--which took time. People mostly just wanted to be on AOL, using IM, email, message boards, and playing Battletech.
feel like this is coming up in a lot of domains at once, but super abstractly: any substantive quality standards are going to result in de facto 'viewpoint discrimination', because certain viewpoints are low quality when judged by those standards. you gotta choose!
Democrats planning to run in November’s midterm elections have been advised not to antagonise pro-AI campaign groups that have amassed more than $300mn to fight for the industry’s priorities. The warnings by top party consultants, corroborated by people close to four different campaigns and party strategists speaking on the condition of anonymity, come despite internal polling for Democrats showing widespread public support for tougher AI rules.
They're calling it the best democracy ever.
www.ft.com/content/7529...
This is more broadly generalizable: If you don't do the research yourself, you have no idea if AI is making stuff up. And if you do the research yourself, then AI isn't saving you much labor, if any.
I’m one of the coauthors of this new SEP entry on early modern rationalism, having truly broken bad from my empiricist roots plato.stanford.edu/entries/rati... #philosophy
We are fast approaching the headline, "Vance says Jesus should stay in his lane, avoid politics."
Man Who Threw Molotov Cocktail At Sam Altman’s Home Claims He Was Following ChatGPT Recipe For Risotto
Man Who Threw Molotov Cocktail At Sam Altman’s Home Claims He Was Following ChatGPT Recipe For Risotto theonion.com/man-who-threw-molotov-co...
"In a few years everyone will be using Facebook phones in India, connected via floating Wi-Fi balloons, spending Libra on Oculus headsets so they can be part of the Metaverse!"
In general, (re)orienting education wholly or chiefly to specific perceived market needs is always going to produce disaster once the near term passes. It also eviscerates institutions that might otherwise help people weather change, intellectually and in other ways
I see there's an analytic vs continental thing blowing up on Substack and: no. No we are not doing this. It is not 1998 anymore. Searle's dead, Derrida's dead, we're not doing this anymore. Go get a sandwich and look at a pigeon.
I don't think anyone cares if two senators date. I think people are troubled by dating subordinates. Even in restaurants, there's usually a rule against relationships with those who answer to you.
Everyone roots for Jim and Pam, but the Michael and Jan story line was an awful mess.
New paper with Olivier Michalon. We ask whether transformers that separate syntax and semantics into parallel streams classify sentences better than single-stream models. They do not. We think this matters for how the autonomy of meaning and grammar should be explained. 1/6
doi.org/10.1080/0952...
"Turns out students aren't great at asking questions well." No fucking shit. Literally any educator could tell you that. 90% of my job as a high school librarian is trying to figure out what kids actually want versus what question they asked.
🧵Excited to announce—
"Reimagining the binding problem(s) for the 21st century": A VSS Symposium
St. Pete Beach
@vssmtg.bsky.social
May 15th, 10:30am
Presenters: Peter Tse, JohnMark Taylor, Seda Karakose-Akbiyik, Ana Chica, Anne Sereno, & Jake Quilty-Dunn
visionsciences.org/symposia/?sy...
AI companies' business model was clearly based on getting kids hooked on AI by forcing it into schools, but now kids just hate AI and are pissed at their schools...
#NihilismAndTechnology
I try not to take things personally and all, but a journal gave me only two weeks for a revise and resubmit. And, like.... It kinda feels like they hate me.
16 GW of power deals announced by 5 hyperscalers in 2026
More evidence that Big Tech is locking us into a new generation of fossil fuel infrastructure and destabilizing our climate: the 5 largest hyperscalers have announced 16GW of power deals this year, and only 3GW have been for clean power.
I'd love to be added. Thanks for making this!