After hacking incidents and the industry's existential risk warnings, we need a regulatory ban on RSI, as competition drives unchecked autonomous acceleration
New York Times Ezra Klein
Related Coverage
- Is this how the world ends? Extinction scenarios are taking over the AI debate. Washington Post · Nitasha Tiku
- AI Companies Are Warning About Frontier Development. An AI Researcher Says Regulation is Key To Avoid ‘Catastrophe.’ International Business Times
- Pacing the frontier may be sincere, but it would also be strategically useful for frontier AI labs to have time to reduce overhangs caused by model advancement Stratechery · Ben Thompson
Discussion
-
@timnitgebru
@timnitgebru
on x
Even after their brief scrutiny after SBF, after the castle, after the sexual harassment articles, what he had to say was write articles saying not to write them off and that they were good actually.
-
@politicalmath
PoIiMath
on x
I consider @ezraklein to be way more educated on AI than a normie & I assume he's talked with lots of AI experts. So it's interesting that he doesn't understand that we are well past the event horizon on this particular strategy.
-
@tyler_m_john
Tyler John
on x
Most of the people I know who work on AI x risk don't have it as a “worldview” as in something they take for granted. It's more like one highly plausible and natural trajectory.
-
@ramez
Ramez Naam
on x
Separate from Ezra's concern, this proposed first step is unworkable. Everyone else gets to use AIs to write code, but not the labs? Won't happen. It's hard for me to find proposals to slow down AI development that don't require surveillance or infringement of basic freedoms. [im…
-
@matthewstoller
Matt Stoller
on x
Barack Obama and Ezra Klein have built careers confirming to dull liberals what they want to believe, paving the way for the even dreamier Jon Ossoff.
-
@krishnanrohit
Rohit
on x
Ezra is in his less wrong era
-
@robertwiblin
Rob Wiblin
on x
Blistering piece from Ezra Klein. Calls for a comprehensive ban on Recursive Self Improvement starting immediately. RSI is a grey area, so how do you define it? Very expansively. Then get firms to apply for narrow exemptions. YouTube link below. https://www.nytimes.com/...
-
@binarybits
Timothy B. Lee
on x
To be clear, I disagree with the worldview here more than the conclusions. I think it's a good idea for the labs to slow down. But I think people are going to find that “RSI” is a chimerical concept akin to “AGI.” Making it the linchpin of regulation isn't going to work well.
-
@binarybits
Timothy B. Lee
on x
This column makes it clear that Ezra Klein, the left's smartest and most influential pundit, has fully bought into the AI X-risk worldview. Those of us who think these ideas are confused have completely failed to make our case. https://www.nytimes.com/...
-
@samarthguptama
Samarth Gupta
on x
The collective problem action between companies and across countries that our government should be solving https://www.nytimes.com/...
-
@sethackerman
Seth Ackerman
on x
i get that people are skeptical. but remember, these AI people have been warning about these freaky sounding scenarios since before they even *started* writing the programs that subsequently acted out the freaky sounding scenarios https://www.nytimes.com/...
-
@alecmacgillis
@alecmacgillis
on x
“Where these labs are, you can't build an 8-story apt building w/o an agonizing public review process, probably not even then. Yet somehow it is possible for these labs to unleash a swarm of 40K A.I. agents to build a society-altering superintelligence w/o so much as a hearing.”
-
@dorotheabaur
Dorothea Baur
on x
“It surprises me that this counts as a radical proposal, but here it is: If you are losing your ability to evaluate the models you have now, maybe don't let them build models you'll be even less capable of controlling in the future” @ezraklein https://www.nytimes.com/...
-
@eoinhiggins_
Eoin Higgins
on x
One of the reasons I'm so pedantic anthropomorphizing about AI “thinking” and acting independently is that when you do that, you end up with what Ezra Klein is saying in one part of his lengthy essay today on the technology. https://www.nytimes.com/...
-
@loomdoop
Y Disassembler
on x
Exra Klein is the absolute last person anyone should listen to about these AI companies. He's been PR guy for the Effective Altruism cult for a decade. He even re-oriented Vox around it.
-
@sbeverts
Steven Everts
on bluesky
Watch this. My three conclusions: — 1. The AI you and me use is very different from the frontier models. — 2. Because of recursive self-improvement, we are very close to loss of control. — 3. We need to control the frontier, not just pace it. — The mother of all collecti…
-
@nathaliemarechal.net
Nathalie Maréchal, PhD
on bluesky
The @nytimes.com is a straight-up LessWrong post. What is even happening 😳 www.nytimes.com/2026/09/20/o...
-
@mjgault
Matthew Gault
on bluesky
I simply refuse to take people at their word when they're trying to sell me something
-
@tribelaw
Laurence H. Tribe
on bluesky
Ezra Klein is right: RSI absolutely must be stopped before it's too late. But here's the question: how can we possibly ENFORCE a ban on AI's recursively building improved AIs even if we and China agree on the terms of such a global ban — a very big “if” in itself? — www.nytime…
-
@viksit
Viksit Gaur
on x
this. alignment risk is not the same thing as who defines what's bad.
-
@firesidealpha
@firesidealpha
on x
Ben Thompson calls Meta's Muse a bearish signal for the frontier labs, because a not-state-of-the-art model now makes a product stickier than any chatbot “The key for the frontier labs, then, is to build those user touchpoints while they have superior capabilities.”
-
@cnakazawa
Christoph Nakazawa
on x
Great read https://stratechery.com/...
-
@rabois
Keith Rabois
on x
Several highly provocative thoughts.