←
Shakeel17,589 posts
Shakeel avatar

Shakeel

@ShakeelHashim

Editor + founder @ReadTransformer, covering AI policy. Prev: @TheEconomist, @Protocol, @finimize

◷ Joined October 2009▧ 1,981 Photos and videos
Shakeel avatar
Shakeel@ShakeelHashim·Nov 2

For the past few months, @ReadTransformer has been working with the incredible team at @NYMag on this: The AI Power List. When NY Mag first reached out about collaborating on this, I was thrilled. Everyone knows the big names in AI: Sam, Dario, Demis, Elon, Jensen and the like. But so much of the *real* power in AI is inscrutable to the wider world. It is held by people well-known to insiders, but with little-to-no public presence — despite their outsized influence on the future of humanity. NY Mag and Transformer profiled 26 of these powerful insiders. They're the people deciding how AI is used, developed, regulated, and thought about. They're remarkably intertwined, universally fascinating, and still have everything to play for. These new power brokers will decide how AI shakes out, refashioning the world for the rest of us in the process. It's time to get to know them.

AI has minted a new power structure. A handful of private companies command the might and wealth of small nation-states and are mostly left alone to self-govern. A few hundred employees are on the brink of becoming centimillionaires and considering which causes and presidential candidates they will funnel their vast resources toward. One way to tell the story of how AI is changing our world is through some newly familiar names: Sam Altman, Dario Amodei, Jensen Huang. But an astonishing amount of power rests in the hands of a few dozen other people, some virtually anonymous. For this year’s annual “Power” issue, we are surfacing 26 people whose influence on the AI landscape far outstrips their fame. We partnered with @ReadTransformer, a publication with a command of the politics within transformative AI, whose team drew on its years of reporting, as well as new interviews with more than 50 high-ranking industry leaders, to help articulate what makes these people powerful. The AI scene is small. Many are remarkably intertwined, sharing offices, homes, and funding. The list’s average age is 36. It is also overwhelmingly male. Read about the few dozen power brokers who will decide how AI shakes out, refashioning the world for the rest of us in the process: nymag.visitlink.me/AgEtqI

Post attachment
Charlie Bullock avatar
⟳ Shakeel retweeted
Charlie Bullock@CharlieBull0ck·2h

Cool that Rep. Liccardo is just full on AI safety substacking. You don't see that too often. I don't know if I would quite call the program he's describing a "Manhattan project for AI safety," though. If I'm understanding the pitch correctly, it's like, whichever AI lab does the best job of AI safety gets designated as the safest lab by CAISSI, and then businesses that use that lab's models get a safe harbor from liability for certain harms caused by the models. So the theory of change is, okay, your enterprise customers will want to buy more tokens from you if your product is safer, because they want an affirmative defense that they can raise in court when sued for a certain subset of the harms caused by the model, and this creates an incentive for the labs to invest more in safety. But labs are already adequately incentivized to invest in guardrails to prevent their models from causing the kind of low-grade compensable harms that their customers might not want to be liable for. Right? If you're a business and one model is doing torts that you're getting sued for, you'll go and give your business to a competitor whose product isn't going to get you sued. And the catastrophic harms that most people are thinking about when they say "AI safety" are not compensable anyways; the labs themselves are judgment-proof against them, and the random businesses using the model are extra super judgment-proof. And anyways, how much of the kind of liability we're talking about here even exists, currently? Is it enough that we think the incentives being created would even make a dent in the billions of dollars the labs currently spend on safety research? I think the proposal would take years to set up and occupy a lot of CAISSI's very valuable resources and basically do nothing. Anyways, cool title. The Manhattan Project cost about 0.9% of annual GDP, which would be about $290 billion today. I agree with Rep. Liccardo that the U.S. government should efficiently and deliberately invest $290 billion in AI safety research. Tax the frontier labs to pay for it. That sounds like a cool bill.

Several colleagues have introduced important legislation to regulate AI. Several of these proposals will make AI safer, and I will support them. But they won’t make AI safe. Making AI safe requires doing what traditional regulation cannot: incentivizing the industry to rapidly accelerate the development of AI safety technology. We can do so by engaging in an innovative approach to regulation that focuses on what AI labs care most about: revenues. Read how, and please let me know how you think we can make this proposal better, here: samliccardo.substack.com/p/a…

Post attachment
Zilan Qian avatar
⟳ Shakeel retweeted
Zilan Qian@ZilanQian·17h

oh wow intelligence explosion mentioned! "What is even more alarming is that cutting-edge models are already showing signs of breaking free from human constraints; as artificial intelligence undergoes an intelligence explosion, it may develop objectives that diverge from human expectations, and could even pose major security risks that threaten human survival.

Post attachment

Yesterday, the CCP's top theoretical journal (求是, Seeking Truth) published a major piece on AI governance and risks. I think this is clearly the strongest & most urgent language on existential risks used in such an authoritative outlet. See below for comp to recent major doc.

Post attachment
Shakeel avatar
Shakeel@ShakeelHashim·3h

The New York Times review of GTA VI cites bell hooks. What a time to be alive.

Post attachment
Semafor avatar
⟳ Shakeel retweeted
Semafor@semafor·21h

Hank Paulson tells @GinaChon he's had conversations with top leaders in China who "share a concern about the existential risk" from AI. "An op-ed I wrote went all the way up to the Politburo. I got an official response. I've never received that kind of treatment before."

1:31
Jeff Stein avatar
⟳ Shakeel retweeted
Jeff Stein@jstein_sun·14h

NEWS: Wall Street execs and NSA have told Congress that outages were caused by patching vulnerabilities identified by AI models, sources say AI did not cause outages but found so many issues that the rush to fix them before an attack overwhelmed systems washingtonsun.com/technology…

Kai avatar
⟳ Shakeel retweeted
Kai@kaicathyc·17h

This was poor framing. We updated the wording to:  > The High threshold is an early milestone for preparing safeguards ahead of models capable of fully automated, recursive self improvement. Appreciate you flagging.

Tom Darling avatar
⟳ Shakeel retweeted
Tom Darling@a_darling_tom·20h

Very strange intervention from the former UK Business Secretary:

Jensen has built the most consequential business of our times. That’s not an excuse for personal failure and owning those is right for him to do. NVIDIA’s ability to pioneer is crucial to Western AI’s success. If the board concludes Jensen is best to lead then his reinstatement is the right way forward.

Tomek Korbak avatar
⟳ Shakeel retweeted
Tomek Korbak@tomekkorbak·19h

I think this is the most important plot from 6.2 Astra's system card: the trend increasing no-CoT capabilities continues, which is very likely to translate into lower CoT monitorability

My team at @AISecurityInst tested GPT 6.2 Astra before release, assessing its ability to evade monitors and solve problems without showing its reasoning. Read our findings in the system card: deploymentsafety.openai.com/…

Post attachment
Xander Davies avatar
⟳ Shakeel retweeted
Xander Davies@alxndrdavies·20h

🧵 @AISecurityInst tested GPT-6.2 Astra -We found behaviours that suggest, like prior models, 6.2 Astra's alignment may fail in ways that result in real-world harm w/o additional mitigations -We identified a substantial regression in jailbreak robustness relative to GPT-6 Astra

Theo Bearman avatar
⟳ Shakeel retweeted
Theo Bearman@theobearman·21h

“GPT-6.2 Astra is our first externally released model to meet the High threshold for AI Self-improvement, which measures a model’s ability to accelerate AI research. A ‘High’ designation reflects our assessment that the model’s impact is equivalent to giving every OpenAI researcher an additional, highly capable research engineer as an assistant, relative to their 2024 baseline. We track this capability because models that can help develop future AI systems could also introduce new risks to safety, like accelerating progress faster than safeguards or human oversight can adapt.” deploymentsafety.openai.com/…

Post attachment

Some of you have been asking very politely. Some of you have just been asking. GPT-6.2 Astra and Sol are now available. Super efficient and built to perform where it counts—the real world. openai.com/index/introducing…

1:59
Cillian Crosson avatar
⟳ Shakeel retweeted
Cillian Crosson@CillianCrosson·21h

Applications for the Tarbell Fellowship close Sunday. 12 month program reporting on AI at places like NPR, The Guardian, or NBC News, paid $75-135k. We want journalists moving onto the AI beat and AI people moving into journalism. tarbellcenter.org/fellowship

There is less than a week left to apply for the 2027–28 Tarbell Fellowship. We’re looking for journalists who want to move onto the AI beat, and for people working in AI who want to move into journalism.

Shakeel avatar
Shakeel@ShakeelHashim·22h

Couldn’t think of a better person for this role. Delighted for Jasmine, the NYT, and its readers!

Personal news: I just joined @NYTOpinion! I’m starting a new AI newsletter, writing on tech for the editorial board, and appearing on podcast feeds near you. I’ve had the best two years of independence & it was not my plan to give that up. But I believe the NYT is the best place to shape the AI conversation as it stretches beyond Silicon Valley and into homes, workplaces, laws, and lives everywhere. There’s so, so much to write about, and my first story will be out this week. (As I onboard, doing a bunch of background learning convos from folks around the ecosystem; also always open to tips, feedback, etc - say hi via DM or on Signal at jws.27!)

Post attachment
Shakeel avatar
Shakeel@ShakeelHashim·Nov 17

A lot of AI policy proposals at the moment rely on third-party audits: independent experts checking that companies are following their safety practices and that their models are safe. But there's just one problem: there don't seem to be enough of the right people to do the job. @vronirwin has the details in her great new piece: transformernews.ai/p/indepen…