Quantumfreakonomics
No bio...
User ID: 324
Quantumfreakonomics
1d ago
This would be one of the most successful astroturfing campaigns of all time. If China was behind it, don’t you think they’d be putting a lot more emphasis on,”you know, all these data center chips come from Taiwan. Why are we subsidizing Taiwan’s defense only for them to suck all of our water?”
Quantumfreakonomics
1d ago
We did have a thread about this a few weeks ago.
A big problem for the AI companies is that, while the main issues with data centers can be solved with more infrastructure buildout (more power generation, more transmission lines, more water plants), it turns out that people don't like infrastructure buildout either!. Its true that the footprint of the data centers themselves are quite small in the grand scheme of rural American land use, but once you start building cross-county high-voltage power lines, gas turbines, gas pipelines, and water lines, you end up stepping on a lot of landowners' toes.
The power issue is not trivial. The 2024 Situational Awareness paper predicted that in a permissive regulatory structure, AI power demand in 2030 would be 100% of total US electricity generation in 2024. We would need to double electricity production in six years to keep up. That breaks things. That generates meaningful pollution. This scenario will almost certainly not play out (for reasons which at this moment must be all too obvious), but it is true than many AI techbros want this to play out.
Quantumfreakonomics
2d ago
I mean, that's a trivial way for what I said to be true, but even the original work is of quality that rounds to zero. Zero is not meaningfully lower than something that rounds to zero.
Quantumfreakonomics
3d ago
The state of Arday’s subfield (sociology of education) is so bad that Arday’s entirely fraudulent work is not of meaningfully lower-quality than the median output.
After leaning more about his work and its context, I would like to retract my previous statement. To paraphrase Bismarck, the integrity of sociology of education research is not worth the bones of a single Pomeranian grenadier.
Quantumfreakonomics
3d ago
I think it’s relevant that this happened in Europe. America had its reckoning with race-studies scholars of questionable quality a few years ago. The threat of being called-up to Capitol Hill for sworn testimony and having your university’s federal funding cut seems to have put the fear of God back into university administrators. They would be scared to engage in this blatant of a cover-up on this side of the pond.
Ghent University sociology professors ought to be careful what they wish for right now.
Quantumfreakonomics
4d ago
But maybe I'm just biased because I can't imagine a pretty feminine woman like Zendaya ever really being a school shooter.
Quantumfreakonomics
5d ago
Someone needs to do an old-school Glenn Beck-style chalkboard conspiracy diagram connecting everyone in the SF effective altruism/AI community. Daniela Amodei (Dario’s sister) is married to GiveWell and Open Philanthropy founder Holden Karnofsky. Im not sure what this means, but I’m sure it means something.
I’m starting to see this as a big L for Singerite utilitarianism. Anthropic and OpenAI were both founded for an ostensibly charitable purpose. Some of this is cynical posturing, but I think most of the key people genuinely feel like they have an obligation to be charitable with the windfall from the technology they are building. A normal company fulfills their social obligation for charity by sponsoring local little league teams, the regional food bank, parks, and schools, but since Anthropic (and to an extent OpenAI) are Singerite utilitarians, they see local feel-good projects like these as “ineffective” altruism.
More options
Context Copy link