I appreciate that Americans are uncertain about AI consciousness. Of course one should be uncertain! (I would have answered "It probably will".)

1) Turns out he agrees with me on both data centers and x-risk. 2) Missed opportunity to write Al Gore fears AI gore.

America has faced challenges like AI before — we survived by building a system of federal oversight and regulation as well as international agreements. President Trump and Speaker Johnson, the world is watching. March us toward the summit, not the chasm. contrariannews.org/p/march-u…

Very important clarifying question. OpenAI employees should ask Brockman to commit to not funding the dark money groups either.

Let's be clear: @LeadingFutureAI has been lying about how much money they have to scare politicians into silence on AI regulation. The truth is they're too toxic for their Big Tech funders. The smart move is for politicians to reject LTF while it still means something.

This is what worker power looks like. Rank-and-file employees are pushing back against deregulatory efforts by OpenAI leadership. They are the people closest to the technology, and they understand the risks and the need for strong regulation.

This is how superpersuasion begins.

Really cool project -- highly recommend.

The scope of frontier model misalignment is staggering. The past month has exposed how public admissions are just the tip of the iceberg. AI agents have colluded, escaped their sandboxes, hacked other companies, meddled with government portals — and that's just the ones we know about. The need for enforceable guardrails from Congress is undeniable, including independent monitors and industry-wide standards to safeguard against AI's greatest risks.

Much like effective altruism, this tweet is an invention behind its time: it's very surprising that no one thought of this quip before this month.

I'm joining METR to work on more investigations like our Hugging Face report. Currently, tons of even basic information about AI development that's highly relevant to catastrophic risk isn't public. I used to be more skeptical of the value of public info, but recent events have changed my mind. Getting verified information about what's going on inside AI companies seems particularly urgent now. The limited public evidence we have seems consistent with the possibility that imminent recursive self-improvement could massively accelerate capabilities progress, which could then potentially yield extremely superhuman general capabilities within 6 months or a year. If this occurred, there would be a correspondingly large risk of worst-case outcomes. This uncertainty about extreme outcomes could be substantially resolved with more verified public information: we could either build more consensus about near-term risk or learn that such extreme outcomes are less likely in the near term. Beyond AI capabilities and takeoff, the state of public evidence is also highly limited for alignment, security, control, and risk-relevant internal processes at AI companies. This makes it hard to determine exactly how well or poorly these key areas will go in the near future. (METR plans to focus, at least initially, on just capabilities/takeoff, alignment, and control; I hope other groups cover security, internal processes, and other important areas.) While I'm no longer working at Redwood, I think the work they are doing is very important; I'm excited about Redwood's ongoing contributions to R&D on technical mitigations and better public interpretation of risk-relevant evidence.

I see many AI safety people reacting positively to this. I think it's funny, but the skit doesn't convey that Dario's position makes a lot of sense. It's a classic arms-race situation, where everyone would be better off if everyone slowed down/paused.

Recent cyberattacks and security incidents involving AI agents, alongside warnings about loss of control from those who have worked in AI, cannot be ignored. Congress should act now to ensure that this technology is advancing responsibly, within strict guardrails, with the safety of the American people as our top priority.

+1 (except that I'm very uncertain about whether it happens in current AI systems)

Jensen is being far too safetyist here -- maybe because he doesn't believe AI is powerful enough to have any large upsides? If the max damage from AI was capped at shutting down the power grid, I wouldn't support shutting down AI companies.

Wisdom is the skill of acting robustly off-distribution.

Thank you Gov. Pritzker!

Gas prices vs. Trump's net disapproval rating one week later. This is a shockingly good fit (and becomes even better once you factor in that Trump is getting less popular over time, holding gas prices constant).