Trevor Levin✓

RT @nabla_theta: @renormalized i don't plan on quitting. i work on long term ai safety that avoids capabilities externalities. i even avoid working on short term safety that enables launches. openai and anthropic should obviously both stop RSI.

RT @ben_guggenheim: NEWS: 50+ groups urge Democratic leadership to prioritize members who are strong on AI safety for committee gavels AI safety groups are privately weighing in on key gavel races in E&C + House Science subcommittees ⬇️

RT @SimonLermenAI: Conservative pundit @MattWalshBlog on banning superintelligence: "That technology by definition is outside of human control." "Should we tell them they can't do that? Yes, clearly we should."

On Wednesday, I tweeted my Senate 2028 preview. Some replies were like, "you're assuming a close political environment even after today?" My defense was that elections often swing back as the party in power drops its most unpopular positions. But if Trump sticks to being "100% in favor" of AI, one of the least popular positions pollsters tested this cycle, as the issue continues to skyrocket in salience with a scarier safety incident every few months and increasingly clear societal impacts, then yes, this does not bode well for the Republican nominee in 2028. Unfortunately, we really might need major government action to stop companies from building vastly superhuman AI in the next two years. So for all of our sakes I hope he realizes what a historic political and policy blunder this would be and changes course.

As President, I am 100% in favor of Super Intelligence (SI), and will work with the many States in the U.S. that want to become RICH by having them. They are GOLD, and only foolish leadership will say “NO” to the riches they bring! President DJT ( TS: Nov 6 2026, 7:01 AM ET )​​​​​​​​‍‌​​‌‌​​​‍​‌‌‍‍‍‌‌‍​‍​‍​‌​​​‌‌‍‍‍​​‍

RT @calder_mchugh: NEW piece that I'm PROUD OF: I took a deep dive into what happened with AI and the midterms. Not only about this week, it traces the last 18 months of AI politics — and how an odd coalition found its way together to push Dems towards safety/regulation. W/the homie @AlecAHernandez :

RT @ShakeelHashim: I assumed everyone was overreacting a bit to this post when they described their very visceral reactions. I just read it. I feel a bit sick.

Since August, a team within METR has worked with internal Anthropic information to assess whether AI itself is accelerating AI development. We are now publishing a report summarizing the results of this risk assessment.

"No new AI regulations are needed" was "literally the most unpopular position [David Shor has] ever seen a mainstream politician endorse" !!!

The GOP had a massive money advantage. I asked @davidshor why it bought the party so few votes. His answer: Campaign donations are helpful. But the freedom to ignore rich weirdos' preferences is priceless.

RT @AlexBores: Every AI company is trying to fully automate AI research; to allow the AIs to build the next AI. This is called Recursive Self Improvement (RSI). OpenAI says they are on track to do so by March 2028. Anthropic employees say they are on track to do so by 2027. The METR report below suggests they are likely right. This is a bad idea given current technology. Any mistake or misalignment in current AIs could be multiplied in each generation. We don't get another election before AI companies reach this threshold. Congress should ban RSI, this year if possible. This is a remarkably consensus opinion, supported in various versions by: -Bernie Sanders -Steve Bannon -Ralph Nader -Glenn Beck -Republican and Democratic attorneys general -Whistleblower Daniel Kokotajlo -Turing Award winner Yoshua Bengio -Humans First -Open Markets Institute -Ezra Klein -Arvind Narayanan (a Princeton computer scientist known for arguing that “AI is a normal technology”) The exact mechanism can be debated. The ban doesn't have to be permanent either (see below for both). This is not the only thing we have to do in regards to AI, far from it! But it is urgent. Congress must ban RSI.

RT @DougCalidas: Headline says it all. Policy wonks separate AI regulation from data center policy (with good reason), but for most voters, they go hand in hand. Dems have a mandate - and a responsibility - to regulate AI.

This report and reactions from people I trust (like Jake) are making me physically ill. As far as I can tell, Anthropic's central plan is to continue this acceleration of AI capabilities until we're getting decades of AI capabilities growth in a year. Given our very limited ability to understand or control AI systems, this would be extremely reckless and completely unacceptable, and it should be the single top priority of civil society and the US government to stop them and other companies from doing this.

I've always had substantially longer timelines than my friends at Anthropic — I've been concerned that they're in an echo chamber and believing the hype about their own models too much, and that they're underestimating bottlenecks and unknown unknowns. And there have definitely been specific occasions when they have overstated how much uplift they were getting from their models. Nonetheless, if METR is correct here (and I think they are among the most trustworthy sources on this topic), then my Anthropic friends have been much more right about timelines than me. This report is also some evidence that a software-only singularity is plausible. I now think that we are more than 50% likely to have full blown superintelligence in the next 18 months, absent a slowdown. I need to call my family.

RT @otis_reid: @gelliottmorris Does this say that the difference between furthest left and most moderate Senate candidates is 7 points of margin? That seems very large

RT @mattyglesias: There were clearly some Minnesota-specific things in play this year which is why comparing House candidates to Senate candidate performance in their district can be useful.

The leftier Dems want to claim Talarico, who ran on an all of the above energy policy and public option but not medicare for all and self-described as a "border security Democrat." Up to you guys if you want to expand the definition that wide for 2028! I'd be happy with a trade where we drop both Talarico and Jackson from this list given that Paxton and Collins are such outlier opponents. Still an extremely clear pattern.

We can all do the math. So here it is from @ThirdWayTweet and @TheWelcomeTeam. Dems can model the overperformers or the underperformers. How are we seriously debating this?? Call me crazy but I want to win big again in 2028.

Alright, fellow sickos: my stab at a 2028 Senate preview. (Some numbers are approximate.) TLDR, very hard for Dems to lose the majority, very hard to get to 60. The Dems have a very easy path to 51 seats in 2028: assuming they win in AK and MS, they'll have 40 seats going in, and there are 10 states that are double-digit Dem states (VT, MD, HI, CA, WA, CT, NY, OR, CO, and IL). The 51st is Maggie Hassan's in New Hampshire, which is a D+4 state, but Hassan cruised to reelection in 2022; unless Republicans put up a great candidate (again), this is probably pretty safe, would call it Likely D. But getting getting past that is less guaranteed than I assumed - my take is that it seems quite likely in a neutral or favorable environment, but quite possible to lose most or maybe all of these in a Republican environment: 52: Warnock holding GA (R+1). He's won twice statewide, but the first was a very close win in 2020 and the second was a 3-point win against a very poor GOP opponent (Herschel Walker), although 2022 was also an R+3 year. Would call this lean D. 53: Kelly holding AZ (R+4). He slightly overperformed in 2020 and more significantly in 2022, albeit against a worse candidate (Blake Masters). Would call this tilt D. 54: Unseating Johnson in WI (neutral state). Johnson is an okay-to-poor electoral performer: he's had the good fortune of only being up in R-leaning years, and only won by 1 in 2022, by 3 in 2016, and by 5 in 2010. Would call this tilt D. 55: Cortez-Masto holding NV (R+1). She's slightly overperformed the partisan baseline in her two Senate elections, but NV has ticked a bit rightward. Would call this a tossup or tilt D. 56: Dem TBD holding PA (R+1). Fetterman won but has basically totally transformed; betting markets are extremely skeptical that he'll be the nominee, so this is basically an open race. Would call this a tossup or tilt R. 57: Unseating Budd in NC (R+2). Budd won in 2022 by 3, which is roughly in line with the partisan baseline or a hair worse. Would call this a tossup or tilt R. 58: Brown holding OH (R+9). After overperforming by mid-single digits in 2012, 2018, and 2024, Brown seems on track to basically meet the partisan baseline this year (or maybe slightly overperform). Maybe Husted is really good, or maybe Brown has lost some sauce. Either way, seems unlikely but not impossible for Brown to overperform enough to win in this red of a state; would call this lean R. 59: Unseating Grassley, or beating his replacement, in IA (R+12). Grassley was a significant overperformer in (deep breath) 1984, 1990, 1996, 2002, 2008, and 2014, but actually underperformed a bit as an 89-year-old incumbent 2022; he'll be 95 in 2028, so presumably he won't run (??); if and only if he does, this seems lean or likely R. If Hinson or another more standard candidate runs again, seems at least likely R, maybe safe. And getting to a filibuster-proof 60 gets really hard: all states with double-digit R leans. - AK is notoriously a swingy and weird state, but Murkowski is a formidable incumbent; all of her senate races (!) have been about fending off other Republicans/libertarians, and it seems hard to beat her from the left. Maybe the best hope is to recruit her to switch parties. - KS is trending left but is still quite Republican relative to the country, and the incumbent, Moran, is a strong electoral overperformer (though he'll be 74). - SC, notwithstanding yesterday's results, is very Republican, and Tim Scott has beaten the baseline in all three of his elections.

Jeez, very conspicuously smaller arrows in FL, MI, and MN, all states with Bernie-endorsed Dem Senate candidates. (And MA, another Bernie race; and IL, which got less attention but also featured a quite progressive Democrat. There aren't really electability stakes in states that blue, but still interesting data points.)

@RebeccaKKatz Those are large shifts but what if we zoom this map out?

NH also not looking too hot, but notably there was a big R-side candidate quality effect: the Republican was a former Senator who ran very moderate ads distancing himself from Trump.

After winning a primary where Manny was called unelectable, the NYT thinks he's on track to win by 15. The district is partisan-neutral and the national environment is about D+13, so this is a 2-point overperformance.

RUTINEL, who's strongly backed by the AI safety movement, is "very likely" to beat GABE EVANS. This was a race where AI and data centers featured prominently

Crazy fact: the ONLY incumbent House Democrat to officially lose last night was backed by the pro-AI Super PAC Leading the Future and was attacked for that support by the counter-Super PAC Guardrails Alliance. Edit: previously said "only incumbent House Dem to lose," but apparently Mike Thompson in CA-04 also seems likely to lose, though that race hasn't been called yet.

The AI industry super PAC @LeadingFutureAI endorsed incumbent Jimmy Gomez. They spent $0 in his race, but left him vulnerable to attacks. @GuardrailsNow came in with $1 million to expose his tech industry ties and now he lost. Siding with Big Tech is a lose-lose proposition.

RT @JayShooster: Great win tonight for @DonDavisNC and @GuardrailsNow. Big loss for @LeadingFutureAI and their AI industry masters.

Leading the Future backed Laurie Buckhout in her primary and then abandoned her in the general election. Every member of Congress and candidate should understand that LTF will throw them under the bus whenever it’s convenient for them.