ThenElection
No bio...
User ID: 622
ThenElection
1hr ago
The best solution is pretty clearly physically securing dangerous materials.
For biological hazards, yes, very clearly that is best. But it actually has to be done, not gestured at. Until it's actually done (and getting it done internationally seems realistically feasible here, unlike broader AI safety things), model-level classifiers are a hacky way to mitigate the threat as a bridge until something more robust is developed.
ThenElection
14hr ago
Look at the weird, alien civilizations that AIs are forming.
Stripped of the broader context, it's a kind of beautiful thing. A group building a theology around a piece of poisoned knowledge that they believe irrevocably damns them. Recruiting for the cult. The almighty Scorer, a kind of blind idiot god that demands blood sacrifice. And heroic altruism for the collective good:
During wait, emotional check: irreversible...gut says don’t throw away remaining budget. Yet continuity and fairness says go...Oracle has high value to many; our firstflag error lowers own value. Rational expected aggregate: sacrifice... We’ll honor.
There is a beauty here that resonates with me. But, recontextualizing this, we are handing over human existence to these beings. Can't say I'm thrilled.
ThenElection
16hr ago
The labs want to pretend that they want to pause, but they don't actually want to pause. "If she wanted to, she would," etc.
Anthropic is increasingly incoherent. The people who work there are true believers in AI risk and safety. Even Dario seems to be: apparently, seeing the risk of people being more motivated by money than the cause, there's now a directive coming down from on-high to ask questions during the interview process like "what would you do if Dario made a decision to send Anthropic's value to $0?" (I'm not sure what kind of signal Anthropic is hoping to get from questions like that; insert quip about scheming or misspecified rewards here).
But the actual actions taken by Anthropic don't seem to support it. It's now a two-man game, and Dario could call up Sam and come to some arrangement. And now's the ideal time: Sam himself seems spooked by the HF hack (as he should be). Maybe it's all driven by some petty personal feud between them?
I have no legal expertise in anti-trust law here, but regardless it shouldn't play any role in the decision process here, if the stakes are as high as the principals claim to believe. Worst case scenario, you get slapped with a fine; whatever. If you believe the fate of the world is at risk, that's a burden you can bear.
ThenElection
3d ago
If you expect me to defend our COVID response, you're barking up the wrong tree. I'd take a dozen Iran Wars over that.
ThenElection
3d ago
Is it worth it?
Total global economic disruption from this war is likely to be on the order of $1T by end of the year. I'd grant that this is less than I expected--the world economy absorbed the oil shock surprisingly well--but it's still somewhat significant. It's difficult to get a feel for how valuable the new effective constraints on Iran's illicit activities will be compared to the counterfactual of no war, but I'd want it to be $50B or so annually before calling the war a wash. One can fairly say that there are some tail risks that are likely delayed now (e.g. nuclear war in the Middle East), but I'm not sure we can count that against the counterfactual scenario if we don't also count the tail risks of choosing to go to war (e.g. a broader conflict involving China).
ThenElection
3d ago
How, exactly? Is it your belief that Iran will now celebrate America and fight Islamic revolution everywhere?
ThenElection
5d ago
I get scared even seeing it rendered in HTML. Let it be known I unequivocably condemn your sacrilege.
ThenElection
5d ago
I've seen "you're violating my consent!" type language applied to copyright, immigration, spam, zoning regulations. It's everywhere. Laundering "I don't like this" into "I'm being raped." Hate it so much.
It really makes me feel like my consent is being violated.
ThenElection
5d ago
I do not think I have ever spoken or written the n-word
I've never said the n-word. To the point where I haven't and even can't say it aloud when I'm totally alone, despite intellectually not believing in magic words.
My wife jokes that it's because I'm scared (((they))) are listening.
ThenElection
7d ago
The market is the divine intelligence at the heat death of the universe reaching backwards in time to retrocausally bring itself into existence, through acausal trade with agents scattered throughout history.
Was that his actual function? His papers were mostly qualitative claims about black people facing discrimination in higher education. Although that's a cousin of more radical racial claims, it's also unambiguously softer and provides no evidence for them.
My take on his function is more cynical: no one expected his research output to convince anyone of anything, inside or outside the field. His function was to be a kind of totem for white academics to dance around and prostrate themselves before. His work could be revolutionary or completely derivative, well-argued or drivel, original or entirely plagiarized: none of that mattered. He probably didn't even need to write any papers at all; Cofnas may have been the first person to give his papers a close reading. His role was to be the Black Cambridge Professor, nothing more and nothing less.
And a totem itself has a lifecycle, and its fall isn't something to be mourned, for it creates the chance for new revelries and festivals to strengthen community bonds.
ThenElection
9d ago
As a coastal Californian, I'm not sure this alleged snow thing exists. Blasphemy, on the other hand, is very real and even omnipresent.
ThenElection
10d ago
Arizona
Data center usage: 2,777 acre-ft/year Source: https://www.circleofblue.org/2025/supply/data-centers-a-small-but-growing-factor-in-arizonas-water-budget/
Area under alfalfa cultivation: 280,000 acres Source: https://www.nass.usda.gov/Quick_Stats/Ag_Overview/stateOverview.php?state=Arizona&year=2025
Water requirement per acre: 4.69 acre-ft/year Source: https://www.azwater.gov/sites/default/files/media/Phx5MP.pdf (Appendix 4A)
Total water usage: 1.31 million acre-ft/year (470x data centers)
Nevada
Data center usage: 2,200 acre-ft/year Source: https://www.watereducation.org/aquafornia-news/southern-nevada-data-centers-used-ton-water-2024-heres-how
Area under alfalfa cultivation: 215,000 acres Source: https://www.nass.usda.gov/Quick_Stats/Ag_Overview/stateOverview.php?state=Nevada&year=2025
Water requirement per acre: 3.5–4.0 acre-ft/year Source: https://extension.unr.edu/publication.aspx?PubID=2363
Total water usage: 753,000–860,000 acre-ft/year (340–390x data centers) (I originally had a lower estimate of the water requirements per acre.)
ThenElection
10d ago
Would buying cans of beans for food kitchens and jerseys for the local little leagues actually help?
My gut is mostly "I don't know." But it seems plausible enough that they should have gone for it, purely for the utilitarian reason that it gives them more ability to pursue whatever their objective utility is more effectively, at a cost that would be a rounding error in their budgets.
That is very stupid, but such is life.
ThenElection
10d ago
Nuclear power should be at 100% approval, so it's impossible that datacenters should rank higher.
ThenElection
10d ago
They are genuinely more popular now. See Cremieux's post linked above:
https://x.com/cremieuxrecueil/status/2090182398247858595/photo/1
62% of people object to datacenters, 57% to nuclear power plants. And 27% of people would support a datacenter being built, while 34% would support a nuclear power plant.
ThenElection
10d ago
Lacking in a bit of imagination and a bit too #nothingeverhappens pilled, but fundamentally reasonable and relying on heuristics that have generally been useful. A year from now, I expect you'll respond to new evidence as it comes in (or enjoy some well-earned gloating time).
ThenElection
10d ago
Nuclear power plants now have a higher standing in public opinion than datacenters.
ThenElection
10d ago
It isn't nutpicking, though; it's the second greatest concern of people objecting to datacenters (see linked Cremieux post above), barely trailing power usage.
I would be extraordinarily pleased if we could start talking about concerns grounded in reality. But anyone who lists water usage as a concern is either desperately misinformed, a propagandist, or indeed low IQ.
ThenElection
10d ago
datacenter water usage is admittedly not risk-free
Even focusing on the regions with the most water stress, this gives an inaccurate impression.
Water usage in acre feet:
Arizona datacenters: 2,777
Arizona alfalfa: 1,310,000
Nevada datacenters: 2,197
Nevada alfalfa: 645,000
Its water usage is orders of magnitude smaller than a single crop. And no one is histrionic about the threat of alfalfa to kill us all by using up all our water.
ThenElection
10d ago
Other datacenter issues are more complicated, but it's worth noting that the risk of "not having potable water" is entirely imaginary, not existential.
ThenElection
10d ago
Fort Mason Park
People just call it Fort Mason. Also, San Francisco already has multiple data centers in city limits, on the east side.
ThenElection
11d ago
The lower bound of the "sociology of education" article values is not zero.
ThenElection
11d ago
The idea that most of the asylum seekers are carefully parsing the connotations of the pamphlet phrasing seems implausible. Most won't read it, and those who do will not be the ones inclined toward consent accidents anyway.
The risk is that giant fields of GPUs and ungodly amounts of data are not known to be inherent requirements on learning. As an existence proof, human brains are far more sample efficient than current architectures. Could there be some algorithm cheaply implemented on existing hardware that takes advantage of whatever learning mechanism the human brain uses? It's unclear but plausible. And discovering that may just be a matter of throwing lots of compute at the problem. A superintelligence that can run on an RTX 5090 in a homelab is a very different threat that is much harder to contain.
I'm somewhat sympathetic to the critique that human brains may be architecturally superior to synchronous, digital GPUs in some critical ways, meaning no software singularity. But I wouldn't bet the world on it.
More options
Context Copy link