👀 They called it theft

SCOREBOARD

Thursday close. S&P up 1.1%, a broad rally that clawed back the day after the Fed's rate hike, with Oracle and Nvidia leading. On the week: Nvidia +0.45%, S&P +0.61%, Nasdaq +1.29%.

ON WATCH

On Thursday a federal court unsealed documents from the New York Times' copyright case against OpenAI and Microsoft, and one line is doing the rounds. In a 2023 internal memo, a Microsoft director of applied science called training AI on other people's writing the "largest theft of labor in human history."

The brief, filed by the Times, says the two firms pulled in paywalled articles, built training sets out of them, and stripped the copyright notices off first. One OpenAI dataset reportedly held 91,692 copies of articles from the Times, the Daily News and the Center for Investigative Reporting.

The man who wrote that line, Microsoft's Brent Hecht, was warning his own employer, not going public. Three years on, a lawyer has turned his private worry into the heart of a case that could decide whether AI has to pay for the work it learned from.

The case against reading too much into it: this is the Times choosing which memo to wave, not a judge's finding, and no court has yet ruled the training illegal. Both firms argue it's fair use, that learning from public text is legal, and neither would comment on the filing.

SPONSORED

The next Apple may already exist. Insider sources say Elon has spent two years building a secret device inside Tesla's facilities — one he claims will be "10x bigger than the largest product in history."

There's just one problem: the company is private, and unless you know Elon personally, you can't buy a single share. That was true until Guardian's research team found three public ticker symbols sitting in the launch supply chain.

Click here to see all 3 tickers, free of charge.

You won't hear these names on CNBC — Wall Street hasn't published a word on the connection. But when the launch hits September 21, that quiet ends.

Some are already calling this the biggest opportunity since AI. For anyone who missed Apple before the iPhone, this may be a second look at that kind of setup.

Get the 3 ticker symbols here.

Unlock the free briefing now.

THE MACHINE

Anthropic published a first-of-its-kind gauge on Thursday: its own Claude models now "lead" 26% of the company's AI research and development, up from under 1% in February. "Lead" means the model runs most of a task end to end from a short prompt while a human keeps watch.

Put plainly, the chatbot is already doing a quarter of the work of designing its own replacement. Anthropic calls the gauge its R&D Automation Index, built on a scale from the research group Epoch AI, and says it will keep publishing it as the models get stronger.

To be fair, 26% isn't a robot running the lab on its own. Claude is supervised on every one of those tasks and still works with staff on more than 90% of the job, and this is Anthropic's own number, from a firm that gains when its AI looks like it's racing ahead.

SOMEONE SMART SAID

Karp said this to CNBC on Thursday, and his point is blunt: the AI labs pushing hardest for government oversight may be after cover from lawsuits, not safety. His own customers, he says, keep finding their private ideas turning up inside chatbots, and that "the competitor next door has all of their output."

It widens today's lead. On Watch was one publisher's court fight; Karp is describing the same grievance from big corporate clients, and a lobbying push he reads as the labs trying to get Washington to make the problem legal. He runs an AI company himself, so he's no neutral referee, but he's pointing straight at the money under the safety talk.

THE CHART

The ten biggest companies in the S&P 500, almost all of them AI names, now make up about 40% of the entire index, a bigger slice than at the peak of the dot-com boom. Own a plain index fund and you increasingly own Nvidia, Microsoft, Apple and a short list of their neighbours, whether you meant to or not.

That's why Thursday's bounce was so broad: when the giants move together, the whole market moves with them. The counter is that these are real, profitable companies with actual earnings, a long way from the story stocks of 1999, so concentration on its own isn't a bubble. Still, the more the market leans on a handful of AI bets, the more a wobble in one of them is a wobble in all of it.

HEADS UP

ON THE HORIZON

MEERKAT ON YOUTUBE

Why Europe Is Collapsing. Meerkat's Finance channel on why the continent's growth engine keeps stalling, and what a weaker Europe means for everyone else. Watch on Meerkat Explains Finance.

THE BURROW

Quick vote. A Microsoft scientist privately called training AI on other people's work the "largest theft of labor in human history." Now that it's public, what's fair?

Trivia. The file-sharing service that set off the music industry's landmark copyright war launched in which year? Answer's at the bottom of the email. No peeking.

ALSO WORTH YOUR INBOX

The daily crypto brief from the same burrow. Bitcoin, the news and the numbers in a five-minute read. Join 95,000+ readers.

HOW DID WE DO?

One tap, and Meerkat listens. How did today's issue land? Hit a button, and add a line if you want to tell Meerkat why.

Two days off the machines now. Meerkat's going to read something a human definitely wrote.

Know a writer who suspects a bot's been reading their stuff? Forward them this.

Meerkat is not telling you to buy anything, sell anything, or short anything. Meerkat is not a financial adviser. Meerkat is a small animal with a whiteboard and a copyright notice it refuses to strip. Please do not sue Meerkat.

TRIVIA ANSWER

1999. Napster launched in 1999, and the copyright war it set off, artists and labels against a piece of software that had swallowed their work, rhymes with the fight the AI labs are in now.