Physicist (noneq stat phys), interested in writing, socialism, history, other stuff to distract me from work. he/him. Math isn't real. Planaria dad.
Stat mech factory
- Tweets 19,916
- Following 920
- Followers 1,022
- Likes 274,213
Summary of thread: Any form of capitalism in the global North is incommensurate with non-exploitative relations betn global North and South.
4
2
31
Right now the only way out of the worsening consequences for the human race (not AI existential risk but AI dependence, increasing warming, techno-fascism) is that the AI bubble bursts and the market crashes, creating space for socialism and the green transition.
We’ve known that an alghorithm was “going to eventually kill everyone” for around 200+ years now. It’s called the profit motive. An unyielding “line go up” eternal growth algorithm that has slipped out of human control for a very long time now. Skynet isn’t killer drones, it’s quarterly earning spreadsheets.
1
1
135
Not only is the AI risk hype and call for slowdown a cynical strategy to regulate away competition, but there's no way they can mutually agree to a slowdown. It's as classic a prisoner's dilemma situation as you can get
This whole voluntary AI development slowdown is complete bullshit.
They just can't support their planned capital expenditure burn rate w/o going bankrupt, and they can't IPO because their books are horrific.
So they want to ban open source models and get a taxpayer bailout.
1
4
271
The reason AI safety isn’t much of a “thing” in China is because they understand they are building a tool, not a religion. US AI researchers come out of the “rationalist” cult, which is inflected with sci-fi nonsense and eschatological overtones.
I asked a friend who is in the Chinese tech scene what folks there think about “pacing the frontier” and she was like “AI safety is not really that much of a *thing* here”
122
1,135
7,721
259,928
The communist manifesto was radical for its time but - taking into account that its a propaganda pamplet - today its just a bland restatement of everything capitalism is capable of. If you think it's bad you're ngmi
i hate it when leftists say "oh you dont understand communism because you havent read marx"
when i have unironically read marx and the entire communist manifesto
i fully understand what your ideology is
3
94
@postdocforever.bsky.social / karanji retweeted
Ok but this already happened, Claude was used to kill 130 children when the US bombed a school in Minab, Iran on Feb 28 2026, and when asked about Anthropic’s role in the attack Amodei said it was “a use case that doesn't even violate our red lines” so maybe follow up on that??
🇺🇸 Anthropic CEO Dario Amodei, on the weight of responsibility on AI leaders' shoulders:
"I’m made fun of on social media because our model’s biology safeguards are too strong, and biology students complain... they make jokes [they] can’t use the model.
I would rather they make fun of me than wake up one day and discover that someone used our model, Claude, to kill a bunch of people.
I don’t want that. So I feel both sides of it. I feel both sides of the responsibility."
Writer: Michael
20
1,430
6,003
274,291
I think we finally have an explanation for why Nate Silver's brain is fried: he's spent the last few years going through thousands of lesswrong posts.
The AI risk people actually do show their work! There are hundreds of working papers, thousands of LessWrong posts, literally entire best-selling books about this.
15
396
LessWrong folks have been researching AI risk for years and all we have right now is severely misaligned AI and zero political will to stop it. Failures and losers the lot of them (ideologically, I mean... they did make a lot of money for themselves and for AI companies)
The AI risk people actually do show their work! There are hundreds of working papers, thousands of LessWrong posts, literally entire best-selling books about this.
2
20
810
The US killed a million Iraqis with sanctions before 2001 and another million Iraqis with war after 2001. 9/11 is only significant as a historical event, one that couldn't have happened to a worse country.
5
76
@postdocforever.bsky.social / karanji retweeted
The cynic in me also wonders: one they can sell this as "AGI", would the incentives to perform rigorous biological testing remain? "AI predicted this would cure disease X" might be enough advertising for people to start taking the drug.
1
5
728
@postdocforever.bsky.social / karanji retweeted
Lots of interesting responses here, one point I saw elsewhere on my feed (might have missed it here): showing superiority in math is relevant to the goal of being adopted and funded by the military-industrial complex
1
1
11
810
Is there a reason OpenAI and Anthropic are focusing on solving abstract mathematical problems, where as Tao says, the results themselves are not the goal, rather than curing cancer, where results matter more than understanding?
60
7
291
28,965
Lots of interesting responses here, one point I saw elsewhere on my feed (might have missed it here): showing superiority in math is relevant to the goal of being adopted and funded by the military-industrial complex
1
1
11
810
@postdocforever.bsky.social / karanji retweeted
So “prices differ from labour-values” is not a discovery against Marx. It is in Marx already! The actual question is whether labour-time still regulates capitalist reproduction at the social level once competition redistributes surplus-value. That is the real debate.
6
6
329
10,238
I love the system of magic in a Wizard of Earthsea. It's non-materialist and non-capitalist in a way that Harry Potter isn't. Harry Potter magic is often indistinguishable from advanced technology.
9
375
Just want to clarify that this is a joke, which nobody got. I know NP does not stand for "not polynomial time". It stands for "Nah, it's Polynomial time" and P = NP is actually trivial
Things will really break loose if they can they prove P (polynomial time) = NP (not polynomial time). Anything could happen, there are no rules anymore
1
6
426
@postdocforever.bsky.social / karanji retweeted
A story I am quite proud of. Noticed @OpenAI's bombshell paper referenced a paper by two Indian scientists. Called them up to find a crazy intersting story about their paper.
Read here - indianexpress.com/article/ci…
5
78
418
16,301