unionize heaven

@unionizeheaven.bsky.social

we must organize against God, the CEO of heaven

Exactly this. And in fact, many models *have* been created this way, and decades of prior machine learning systems *were* created this way, consensually. These things are choices! People are STILL working this way! It’s not just that a better world is possible — it already EXISTS. We can nurture it.

John Scalzi@scalzi.com · 2w ago

This is your occasional reminder that generative "AI" in particular could have been developed on public domain materials and/or licensed and compensated works but the major players intentionally, methodically and systemically chose to steal labor from working creative folks instead

It seems quaint now that we do all our banking and race science there, but, nevertheless, it's true: at one time, people mostly wrote jokes on Twitter. 10 years ago, I wrote one about an old asshole and it kinda changed my life. Not his, though. www.cnn.com/2026/10/01/u...

The tweet that defined a decade of Trump | CNN

It’s been 10 years since Jesse Farrar accidentally defined a political era with one pithy, prescient tweet. He wishes it never happened.

cnn.com

something I feel it's important for people to understand about tech news: everyone is afraid of saying "no, that can't happen" because it's hard to prove and you lose credibility if later it does but no one is afraid of saying "that ridiculous thing could be true!" because it sounds open-minded

Timnit Gebru@timnitgebru.blacksky.app · last mo.

This is almost precisely what Emily was saying when OpenAI had claimed that got-2 was "too dangerous to release" circa 2019-2020 so that you understand how long we've had to fight the press performing propaganda-as-news services.

i believe our problems are fixable. we did manage to create a period of time when people could work a job, buy a house, and retire. i don't think this is some crazy unattainable thing, and i do believe essentially all of our big economic problems are fundamentally due to runaway wealth inequality

This is a really interesting read paired with that AEJ paper I shared at some point this week. Both show no clear evidence that faculty affect students' partisanship. Again, no priors will be updated. (I'll find my post with the other paper and reply with it below)

Annaliese Paulson@annaliesep.bsky.social · 3mo ago

New Working Paper with Micah Baum and Joaquin Endara: Students tend to become more liberal during college. We study whether individual faculty make their students more liberal. We find no evidence that faculty affect student partisanship. Working Paper: edworkingpapers.com/ai26-1538 1/

When a chatbot gets something wrong, it’s not because it made an error. It’s because on that roll of the dice, it happened to string together a group of words that, when read by a human, represents something false. But it was working entirely as designed. It was supposed to make a sentence & it did.

Chatbots — LLMs — do not know facts and are not designed to be able to accurately answer factual questions. They are designed to find and mimic patterns of words, probabilistically. When they’re “right” it’s because correct things are often written down, so those patterns are frequent. That’s all.

In the latest assault on First Amendment rights, Columbia University has expelled and fired Grant Miner, President of UAW Local 2710, which represents thousands of Columbia student workers. The firing comes one day before contract negotiations were set to open with the University.

the OTHER thing about the white house’s anti-diversity push is that like old fashioned segregation it is explicitly an effort to undermine unions! like c’mon! don’t be stupid!