Good morning. Andrew here. The S.E.C. is reportedly considering a plan that could rewrite the rules for trading shares in companies. It would allow the creation and trading of tokens that represent a share or a fraction of a share, but may not be authorized or backed by the company the token ostensibly represents, according to Bloomberg. (Which means the token holder may not have voting rights or receive dividends.) Watch this space: It could become a heated battle between corporate America and the crypto community. Also, thanks for all of your notes regarding President Trump’s 3,700 trades. I’d love to hear your thoughts on what trading policy for federal officials should look like. For example, should trading be banned entirely? Restricted to specific windows? Managed through blind trusts? Furthermore, should public disclosures be immediate or delayed, and how should the policy address family members? (Was this newsletter forwarded to you? Sign up here.)
Speaking out about cybersecurity risksConcerns about the power of Mythos, Anthropic’s new artificial intelligence model, have rattled corporate boardrooms and the political halls of power. But since its release, another big worry has arisen: What about those who don’t have access to the tool? While Anthropic made Mythos available to several dozen organizations — including major banks like JPMorgan Chase, Bank of America and Goldman Sachs — to study its dangers, many smaller and international lenders weren’t given access. That raised questions about whether those in Project Glasswing, the Mythos access program, could share information about cybersecurity vulnerabilities with those outside of it, DealBook reports. Anthropic is clarifying what Glasswing participants can share publicly. It began telling program members early last week that they can disclose any discoveries, best practices or code they’ve developed via the model to other companies, industry bodies, regulators and the media, according to a person with knowledge of the matter. “While there was never a specific Glasswing N.D.A., confidentiality protections were something partners asked for at the outset and were built into agreements partners signed,” an Anthropic spokesperson told DealBook. “As the program has matured, we’ve adapted them to ensure this information can be shared broadly for maximum defensive impact.” Concern about information-sharing had grown in recent weeks. Some Glasswing participants had felt constrained by confidentiality agreements, people with knowledge of the matter told DealBook. The terms weren’t created specifically for Glasswing, but did set limits on what these companies could disclose publicly. That was a change from the norm for many lenders: Banks traditionally share information about cybersecurity risks they’ve uncovered. Banking authorities, especially abroad, have been worried about the safety of their lenders:
Industry groups have been coordinating on Mythos. Members of the Financial Service Sector Coordinating Council, including both large and small lenders, are working together on the issue, according to a person with knowledge of the matter. And Anthropic is now a member of the Analysis and Resilience Center for Systemic Risk, a nonprofit founded by big banks to address industrywide security risks. Meanwhile, other A.I. companies are trying to make inroads on cybersecurity. OpenAI recently released GPT-5.5-Cyber, a model it’s also pitching as a next-generation cybersecurity defense. Access to it is also limited to a select group of institutions, though OpenAI recently said it’s widening distribution. Mistral, a French A.I. start-up, is pitching European banks on its counterpart to Mythos, according to Bloomberg, though it’s unclear when that will be made available.
President Trump withdraws his threat to attack Iran today. The president said on social media yesterday that he would hold off on striking Iran again to leave room for “serious negotiations.” That statement, and Washington’s decision to extend its waiver on the sale of Russian oil, have calmed the international oil market this morning. But the average price of gasoline in the U.S. rose to $4.53 a gallon today, according to AAA. High energy prices have spooked the bond market, and driven up the odds of a Fed interest rate increase this year. The Justice Department sets aside $1.8 billion to potentially fund Trump’s allies. The agency said the fund’s aim was to compensate people who claim they were targeted by the Biden administration and Democrats; critics assailed it as a way of funneling taxpayer money to Trump supporters. (It was announced after the president dropped a $10 billion suit against the I.R.S. seeking damages for the leaking of his tax returns.) The Treasury Department’s top lawyer reportedly resigned yesterday. Zohran Mamdani meets with Wall Street leaders. The New York City mayor met with Jamie Dimon of JPMorgan Chase and David Solomon of Goldman Sachs yesterday, The Times reports. (He met with Jon Gray, the president of Blackstone, last week.) Mamdani appears to be trying to mend fences with the city’s business community as he pushes for higher taxes. The mayor had also been criticized by Ken Griffin of Citadel for calling him out by name in a video hailing a tax on pricey second homes.
Musk’s next moveIt took a jury under two hours to dismiss Elon Musk’s $150 billion lawsuit against OpenAI yesterday. Its verdict was swift but narrow, finding that the billionaire hadn’t taken action within the statute of limitations. Both Musk and his legal team made it clear that he will appeal. His path forward may be narrow, Niko Gallogly reports, but the reason that he wants to keep fighting is clear. Musk insists OpenAI swindled him and is a danger to humanity. Both claims are tied to OpenAI’s decision to transform itself from a nonprofit to a for-profit entity. That, he says, incentivizes the artificial intelligence lab to put sales over technological guardrails. He also attacked two fellow OpenAI founders, Sam Altman and Greg Brockman, accusing them of “stealing a charity” and unjustly enriching themselves. Musk’s appeal faces an uphill battle. His lawyers will need to prove there was a significant legal or procedural defect in the trial, like improper jury instructions or evidentiary rulings. Judge Yvonne Gonzalez Rogers herself questioned Musk’s chances. “There’s a substantial amount of evidence to support the jury’s finding, which is why I was prepared to dismiss on the spot,” she said in court after the verdict. But there’s a strong business reason for Musk to press the fight. OpenAI appears to have dodged the serious punishments Musk had sought from the trial, including $150 billion in damages and Altman’s ouster as C.E.O. For now, OpenAI is free to continue its aggressive growth strategy, as it competes with rivals like Anthropic and Google and prepares for an initial public offering. That’s potentially bad news for xAI, Musk’s own A.I. lab, which is so far an also-ran that recently lost more than 50 researchers and engineers, according to The Information. And its biggest recent deal was selling significant amounts of computing capacity to Anthropic. Prolonging the fight creates at least a little uncertainty for OpenAI. And Musk scored a win in some ways by raising questions about Altman’s credibility. Even a slim chance of landing a bigger blow against OpenAI is probably worth it for the billionaire. (And it’s not like Musk can’t afford the legal fees.) “Musk is someone who doesn’t like to lose,” Dorothy Lund, a Columbia University law professor who specializes in corporate law, told DealBook. The A.I. star betting against much of his industryWorries about an artificial intelligence bubble are back, ahead of the latest earnings report by the chip giant Nvidia tomorrow. Declining tech stocks are weighing on S&P 500 futures this morning. The debate was rekindled yesterday following a new regulatory disclosure filing by Situational Awareness, a hedge fund started by Leopold Aschenbrenner, a former OpenAI researcher. Aschenbrenner’s research and views on the A.I. sector are closely followed on Wall Street, and they created a stir again. His hedge fund’s latest 13F filing shows a number of bets against chip stocks — including Nvidia — but bullish trades on companies tied to the actual build-out of data centers. Here are some of the bearish wagers by Situational Awareness:
“It’s not cost cutting; it’s replacing in some cases lower-value human capital.”— Bill Winters, the C.E.O. of Standard Chartered, as the bank unveiled a plan today to cut more than 15 percent of its support staff by 2030 by increasing its use of artificial intelligence.
A Golden Rule for U.S. debtEdmund Phelps, who died on Friday at 92, won the 2006 Nobel Memorial Prize in Economic Sciences for showing in the 1960s that there is no lasting trade-off between inflation and unemployment. Policymakers who try to lower joblessness below its sustainable level by cutting interest rates will get only higher prices in the end, he found. Milton Friedman arrived at the same conclusion independently around the same time. Paul Volcker, a former Fed chair, didn’t cite either economist near the start of his war on stagflation, but he did adopt their logic. In a 1979 speech to bankers, Volcker said: “We can no longer blithely assume we can ‘buy’ prosperity with a little more inflation.” Less remembered today, but highly relevant to current concerns about deficits and debt, is Phelps’s Golden Rule of capital accumulation, writes Peter Coy. The rule was the subject of a paper published in 1961, and refers, of course, to the ancient principle, “Do unto others as you would have them do unto you.” Phelps’s idea was deceptively simple: An economy can save too much, so the current generation skimps unnecessarily. Or it can save too little, harming future generations through underinvestment. The Golden Rule strikes a balance. It specifies a rate of saving that maximizes consumption per person across generations. It treats every generation as an equal claimant. America doesn’t seem to be following the Golden Rule these days. Federal debt held by the public recently surpassed 100 percent of gross domestic product, and interest costs have surpassed defense spending. The lasting value of the Golden Rule is that it turned a moral principle — that we owe something to future generations — into a piece of formal economics, said Laurence Kotlikoff, an economist at Boston University. We hope you’ve enjoyed this newsletter, which is made possible through subscriber support. Subscribe to The New York Times.
Deals
Technology and artificial intelligence
Best of the rest
|
