Good morning. Andrew here. In all of the back and forth over A.I. safety, a new concern has arisen. The chief technology officer of Palantir, Shyam Sankar, argues in an essay that the effective altruism movement is behind the freakout by A.I. researchers. Sankar links effective altruism to Sam Bankman-Fried, the disgraced founder of FTX, who was an adherent. (It doesn’t explain the concerns of Elon Musk or Sam Altman, neither of whom is publicly tied to the philosophy; Dario Amodei has historical ties to the movement but says he’s not a member.) But this argument has been quietly swirling around Silicon Valley, especially among A.I. maximalists who favor less regulation and a need to compete against China. Remember that Palantir is a major military contractor. Also: Our scoop yesterday about Jay Carney’s new job, working with Democrats on tech regulation, has caused quite a lot of discussion online. Barack Obama, whom Carney worked for as a press secretary, linked to a post featuring our newsletter. “Whether this technology results in amazing breakthroughs in medicine, energy and education or unleashes huge economic disruptions, greater inequality, and potential catastrophe will depend on the choices that we make right now — choices that should be made not just by the companies involved, but by all of us,” Obama wrote. (Was this newsletter forwarded to you? Sign up here.)
The messy politics of A.I. regulationOpponents to potential new artificial intelligence regulation are getting more outspoken. In one camp are those like President Trump, who see any slowdown as a weakening of U.S. national security. Another group is concerned with a different issue — and has drawn some unexpected allies. Trump made clear that he opposes any sort of A.I. slowdown, further inserting himself into a debate over whether the technology is advancing at a dangerous pace.
Trump also reiterated his belief that the U.S. needed to maintain its lead over China, and argued that Beijing wasn’t seeking to restrain its A.I. industry. (That said, China’s spy chief argues that the technology could undermine the Chinese Communist Party’s hold on power and called for more regulation.) Other skeptics of the A.I. industry’s latest call are making a different argument, one that has made strange bedfellows, including Lina Khan, the former F.T.C. chair who was frequently criticized by Silicon Valley executives for her approach to antitrust, and David Sacks, the investor and former Trump administration A.I. czar. In an X post on Saturday, Khan argued that the government already had powerful tools to regulate A.I. companies: Shipping flawed AI tools without implementing adequate measures to detect and stop rogue or defective AI agents can be an “unfair or deceptive” act or practice under the FTC Act (and analogous state laws). And some state AGs are already exploring holding AI firms and their CEOs criminally liable when their models participate in criminal activity. Sacks told Bloomberg Television yesterday that he agreed with Khan, and dismissed A.I. companies’ requests for legal cover from antitrust laws to cooperate on safety standards. “You have a basic obligation to make your products safe,” he said. “Do not demand antitrust waivers or liability waivers in exchange for that.” That said, Khan isn’t embracing his endorsement. In a statement to DealBook’s Niko Gallogly, she said: While I’m glad to see a broad coalition for enforcing existing laws against these firms and their C.E.O.s, it is puzzling to see from people who’ve been cheering the dismantling of the administrative state and the corrupting of law enforcement. Who do they expect to do the work? Sacks told DealBook in a statement: I agree with Lina Khan that AI companies face massive product liability exposure if they release unsafe products, even without any new regulatory regimes. In my view, existing product liability laws are a good way to align the incentives of the major A.I. companies with safety objectives. More on the politics of A.I. limits: Today, Steve Bannon, conservative commentator, and Senator Bernie Sanders, independent of Vermont, are set to speak at an event calling for stricter A.I. regulation. And Gov. Ron DeSantis of Florida, a Republican, argued that opposing greater regulation of A.I. was a “losing position” for his party in the midterms.
Oil rallies again. Brent crude, the international benchmark for oil, briefly surpassed $108 a barrel today. A big fear is that a key Saudi Arabian pipeline could be offline for up to five weeks after a drone attack last week, The Associated Press reported, citing unnamed sources. Oil executives, including Mike Wirth of Chevron, are warning of a fuel crisis, contrasting with a rosier Trump administration outlook. Cryptocurrencies slump before a key procedural congressional vote. Traders appear worried that the Clarity Act, a bill to regulate the crypto industry that President Trump supports, will fail to reach a crucial 60-vote threshold in the Senate today. Bitcoin fell nearly 2 percent in choppy trading today as Democrats remain largely opposed to the measure. The Supreme Court rejects the Trump administration’s plan to restrict voting by mail. The ruling appears to draw a line under a weekslong legal tussle that voting rights groups and state officials worried could throw mail-in voting into chaos. But some observers are concerned that the overhaul efforts could still complicate the vote-collection process as Trump apparently seeks to cast doubt on the integrity of the vote in midterm elections. A.I. market jittersThe artificial intelligence boom has driven the nearly four-year bull market. But the prospect of an A.I. spending pullback is beginning to test investors’ nerves. Wall Street is not saying that the A.I. rally is doomed. But worries about the technology, along with concerns about the war in Iran, inflation and potential rate hikes, have some analysts predicting more meager stock market gains ahead. The latest:
The yield on the 10-year Treasury note briefly hit a multi-decade high this morning, pushing up borrowing costs for businesses and households. What to watch: what Treasury Secretary Scott Bessent has to say about the turmoil in the bond market — and the agency’s ineffective effort so far to calm it — when he testifies today before the House Committee on Financial Services.
Interest rates remain in focus, too. Trading this morning indicated a roughly 92 percent odds of the central bank raising its benchmark lending rate tomorrow. That could put Kevin Warsh, the Fed chairman, in direct confrontation with President Trump, who reiterated over the weekend that the U.S. should have the world’s lowest interest rates. The Fed may be the best suited to restore calm to the bond market, some analysts think. How? By raising rates to show the markets that it’s serious about fighting inflation. What’s next? Investors will probably pore over the release tomorrow of the Fed’s so-called dot plot projection on inflation and interest rates for clues about whether the central bank is gearing up for a series of rate increases. That, too, could weigh on the A.I. trade. QUOTE OF THE DAY “The United States has on-orbit space control weapons capable of defending the Joint Force against hostile adversary action.”Air Force Secretary Troy Meink, in the first public acknowledgment that the U.S. has stationed weapons in space. The statement drew consternation from Beijing: “We urge the U.S. side to stop expanding its military force and preparing for war in outer space,” Guo Jiakun, a spokesman for China’s Foreign Ministry, said at a regular news briefing today. Safeguarding data from A.I.Companies worried about protecting themselves from the threat of cyberattacks powered by artificial intelligence are now facing a conundrum. Cutting-edge A.I. models may provide the most powerful defense tools, but what if they take your data while guarding it? That concern was underscored by a report published yesterday in The Information that Palantir Technologies, Booz Allen Hamilton and Nvidia are pulling back or restricting the use of Fable, the latest version of Anthropic’s Claude A.I. model. But Rubrik, a cybersecurity company, is betting on the upside of harnessing the technology. Today, it announced Rubrik Code Guardian, a product that it developed with Anthropic, Sri Muppidi reports. Some companies are wary of using Fable. That’s because it includes a 30-day data retention policy for “safety monitoring by default,” ostensibly to scan for breaches. This has drawn pushback from Anthropic customers and partners, including Alex Karp of Palantir. “We were limited in our ability to actually roll out Fable,” said Eric Simons, the C.E.O. of StackBlitz, a coding platform that uses open-weight and closed-source models for its A.I. coding tool. That, he said, was because many of StackBlitz’s enterprise customers didn’t want it. The solution? “You need zero data retention” for these models, Simons said. Elsewhere, Rubrik sees an opportunity. The 12-year-old publicly traded company has a market value of more than $20 billion and specializes in backing up and recovering customer data. How it works: Rubik’s product creates a copy of customers’ code and identifies possible cyberattack vulnerabilities to developers using Anthropic’s Mythos model. As more businesses adopt A.I., models and agents increasingly have access to sensitive corporate data. Code Guardian is intended to safeguard it, Bipul Sinha, Rubrik’s C.E.O., told DealBook. We hope you’ve enjoyed this newsletter, which is made possible through subscriber support. Subscribe to The New York Times.
Deals
Politics, policy and regulation
Best of the rest
|
