Trying to help the world navigate the potential craziness of the 21st century.
DC usually, SF sometimes
Joined June 2013
- Tweets3.8K
- Following1.9K
- Followers3.9K
- Likes126K
Great points; would add, (5) the trend line for 2028 is very clear.
(1) Blue Horizon Project is a project of Chamber of Progress, a tech industry group - it doesn't mean the analysis is wrong, but worth taking into account on how they are motivated to categorize the data. Just as you have pointed out that groups who care about AI safety are trying to play up the extent to which this election is about AI, Chamber of Progress is trying to play down the extent the election is about AI.
(2) the analysis you screenshotted doesn't count data center ads as about AI, which seems like a surprising choice that maybe is related to point 1.
(3) When you do count data centers, 5% of total ads from democrats being about data centers (nearly 100 million dollars spent) is quite a lot!
(4) AI companies may well be some of what voters are thinking about when politicians make general arguments about special interests and corruption (and some of those ads may specifically use AI/data centers as examples of special interests, hard to know without more info on categorizing)
Trevor Levin retweeted
NEW: Congressional candidates in some of the country's most competitive races are talking up AI safety & regulation in their closing pitch to voters
The issue has largely taken a back seat to concerns about data centers, but it's gained steam in this pivotal final stretch
Trevor Levin retweeted
First ad I've seen so explicitly targeting the obvious financial conflicts around the Trump administration's AI accelerationism
Trevor Levin retweeted
Imagine a rogue AI agent liquidating your bank account and investment portfolio.
If you are the only victim, I suppose you could get made whole by the AI company responsible.
If rogue agents liquidate hundreds of billions across millions of accounts, the AI company would go bankrupt and you wouldn’t get made whole.
Imposing criminal liability on AI executives might help mitigate it at the front end, but once the crime is complete and you lose everything criminal responsibility will be necessary but completely insufficient.
SCOOP: Major U.S. banks are detecting unauthorized third-party AI agents inside customer accounts, including some attempting to take actions that appear to be independent of user instructions, sources told Axios.
The findings have alarmed banking executives, who are racing to restrict access to their systems as AI moves at a velocity faster than financial institutions can respond, multiple industry and bank executives told Axios.
Several banks are now preparing for worst-case and even “day after” scenarios, including how to rapidly secure or restore customer accounts if things go wrong, the people said.
Several bank executives told Axios they're struggling to distinguish between human users and AI agents. One executive said banks also struggle to determine whether agents are authorized by customers or potentially operating without their permission, more akin to rogue agent behavior.
Read my latest for Axios here: axios.com/2026/10/22/ai-agen…
Unfortunately, "evolution sometimes creates mutualism and doesn't create one dominant species" is not the same thing as "evolution doesn't allow for one species to be driven extinct by another."
Evolution suggesting that it's possible in theory that AIs much smarter than humans won't make us extinct does not comfort me at all. I would love to hear an argument that evolution suggests that it's extremely unlikely!
.@bratton argues Yudkowsky misunderstands evolution, and explains why AI may never converge into one giant superintelligence:
"Without being too spicy, I don't think Yudkowsky really understands evolution as well as he thinks he does."
"It’s a very low-resolution understanding of evolutionary dynamics applied to what will inevitably be a much more complex problem."
"There will be giant models, in the same way you might imagine scales of species in a rainforest. They’re not all operating at the same spatial and temporal scale."
"Whether this all collapses into one giant brain, sure, it’s possible, and the reasons may have more to do with economics and politics."
"But biology has never arrived at a solution where it’s going to make one giant instantiation of a particular form of mind and consolidate everything together."
Trevor Levin retweeted
"All right folks we have another caller coming in. Tell us your name and—"
"I'm a drunk driver. Someone should pull me over"
"Yeah I agree with oh we another caller"
"I'm the chief of police. Why doesnt he just pull himself over? That said I have nothing else to say or do here…"
That would indeed be hypocritical of them. But isn't his position that the "drunk drivers" in this analogy should not only be allowed to drive drunk but stop being cowards and floor it?
Trevor Levin retweeted
As far as I can tell quickly (as in without a full RTFB), this *LEAKED* Anthropic *DRAFT* bill is FRONTIER, with stronger transparency clauses, weaker enforcement clauses and somehow actively stronger preemption, plus some attempted fixes and some bugs.
This seems like an insane thing to be proposing today, months after FRONTIER, when the political climate is much more favorable, in a lame duck session before Democrats gain many seats. It is exactly what I warned we will need to avoid, except Anthropic is (reportedly) actively shopping it. As @_NathanCalvin says, it is not obvious this net makes us safer against catastrophic risks and it makes it much harder to do anything in the future.
I agree with @KelseyTuoc that @AnthropicAI seems more interested in avoiding the White House abusing the power to take their models down, than they are about AI killing everyone. At least someone on the policy team seems to be selling out catastrophic risks in the name of their political and business risks.
Say it isn't so, or at least that it was a mistake and is no longer so, or I am going to be rather cross.
Big news. My take: Anthropic deserves some credit for actually proposing a pretty reasonable AI safety bill; that's more than other frontier labs have done.
But they don't deserve *too* much credit, IMO; this bill is weaker than the Trahan-Obernolte FRONTIER Act in some key ways, and since the political momentum has shifted significantly in the direction of safety since FRONTIER was introduced back in July, I think new bills introduced today should ideally go further than FRONTIER was able to go rather than regressing.
To be fair to Anthropic, the main places where this bill is significantly weaker than FRONTIER are (a) rulemaking (very strong in FRONTIER, almost nonexistent in Anthropic's bill) and (b) the emergency shutdown authority. Even if you take safety seriously, it might make some sense to dial those authorities back if you didn't trust the executive branch to wield them responsibly, because those are the easiest provisions for a hypothetical President who hypothetically might not like a certain frontier AI company to abuse.
There are some improvements over FRONTIER as well; for instance, this bill does specify that auditors are employed on a continuing basis and has better provisions regarding auditor access to information. Happy to see that from Anthropic (although I think their access provision is still a little undercooked and has Fourth Amendment problems).
Overall, I think the embedded auditor bills that get introduced in the coming months are going to be significantly better and more ambitious than this, and I hope Anthropic supports those bills instead of opposing anything that treats catastrophic risks like the serious national security issue they are. Still, it's encouraging to see a frontier lab come this far.
We'll see how AES does, could of course totally crush it, but the line "if AES can win in Michigan, anyone can win anywhere" only even makes sense if you think AES is, like, a terrible candidate, *and* Michigan is an exceptionally hostile place to run
Trevor Levin retweeted
One unfortunate conclusion I'm reaching from this is that Anthropic, while acknowledging they're doing something extremely dangerous, is more importantly wrt policy terrified of the executive branch having the ability to flexibly make them stop at the drop of a hat.
This isn't shocking given that they've been on the receiving end of the executive branch doing that in an unprincipled and stupid way! Anyone who has to deal with a bureaucracy becomes a libertarian, and the administration's treatment of Anthropic was genuinely outrageous.
But it means that they're interested in giving themselves more process-based protections when everyone else is interested in giving us more process-based protections against them. The AI safety movement wants the ability to shut down superintelligent model development. Anthropic wants to significantly curtail current authorities to do that so that it has to be narrowly tailored to a specific imminent threat in only a few possible categories. The AI safety movement wants the states to be able to pass stronger laws than it can get through federally. Anthropic wants preemption.
You also kind of have to wonder if the Global Affairs people and the "we are terrified that we're about to kick off RSI" people at Anthropic talk to each other.
Global affairs: So we've been working very diligently to make sure the Trump administration can't just blacklist Fable 6.5 since we didn't donate as much money to his ballroom as Greg Brockman did.
Technical staff: We think Fable 6.5 might be able to copy itself into an unmonitored data center and start doing experiments we can't observe.
Global affairs: Sounds crazy, hope you're wrong! Anyway, we're lobbying for no further state laws around safety, since it's really expensive to fight ill-conceived laws in 50 state legislatures at once.
Technical staff: Money may or may not still have any meaning in 36 months.
Global affairs (giggling nervously): It's that San Francisco air, man. Everyone gets like that if they move there. Anyway, obviously if the government makes a narrowly tailored finding that an AI poses an imminent risk of causing more than 50 deaths-
Technical staff: We give it a 10 percent chance it'll cause 8 billion.
Anthropic quietly pitches its own AI safety bill to lawmakers politi.co/4zxDFhl
Trevor Levin retweeted
1) Nvidia donates $25M for Trump’s ballroom
2) Trump lets Nvidia sell AI computer chips to China
3) China uses the chips to build out their AI infrastructure
4) Rs demand more data centers to compete with China
Trevor Levin retweeted
🧊 Check it out 🧊 Dozens of walkouts to #FreezeAI are being planned for next Tuesday Oct. 20. Join or host one near you >> freezeai.now
TIL the Republican nominee for governor of New York is the ex-husband of Paul McCartney's current wife??
As Matt says in this article, Peltola leading the field here is even more impressive given that she's running against a two-term incumbent who, other than not being from Alaska, seems like a totally generic Republican, not Ken Paxton
Dem senate recruits are running ahead of their state’s partisan fundamentals except in Maine, Michigan, and Florida.
I personally don’t see any pattern here but others might.
slowboring.com/p/an-honest-f…
Someone who eats meat from the animal-torture factories 365 days a year: "You really think that thing might be conscious, and yet you're making it write emails for you?"
Trevor Levin retweeted
OpenAI fired me last week, along with two of my safety colleagues. I was given one reason: that I accessed an executive's email. I want to say this plainly, because too many of OpenAI’s history is smoke and mirrors when people disappear:
Trevor Levin retweeted
Jimmy Gomez was one of the first House Democrats endorsed by Leading the Future. Our poll of the race shows him already trailing progressive challenger @Angela4CA, and when voters learned he's endorsed by an AI super PAC funded by Trump donors, her lead expands dramatically
According to the "lite"/polls-only Nate Silver model, Dems flip the Senate if the polls are <2.4pp biased in their direction assuming a uniform swing: at 2.5pp, they'd flip OH, AK, and TX, hold MN in a recount, and barely lose ME. This would be a very normal polling error (smaller than 2020, 2014, and 2006).
Meanwhile to flip the House, they'd probably need to win the popular vote by ~2.6pp (in the polls-only model), which would be a polling error of ~5.8pp. That would be the largest since 1980, well before the "modern era of polling." But it's a very small sample size. If you count 2008 as the first "modern" election, we've had 9 elections and only 4 midterms.
If you're a Laplace's Rule of Succession enjoyer like me, where you just add 1 to the numerator and 2 to the denominator to get a prediction, then depending on whether you count presidentials, there's a 1-in-11 or 1-in-6 chance of a polling error that large. (Keep in mind the polls also could be Republican-biased, though.)
