Good morning. I popped into a panel discussion about AI (what else?) at Fortune’s Most Powerful Women Summit yesterday afternoon and overheard several intriguing things.
One that stuck with me? What AI won’t touch. According to one tech executive: relationships, situational and contextual judgment, innovation and creativity.
AI can help, the exec said, but humans have to “invent the future.” Well said. —Andrew Nusca
Want to send thoughts or suggestions to Data Sheet? Drop a line here.
Finally, a new Apple iPad Mini

The iPad Mini is Apple's most divisive product. Some people love it. Others can't understand why anyone would want a tablet only slightly larger than an iPhone Max.
But Mini-philes got a surprise bit of good news Tuesday when Apple unexpectedly announced the first new version of the device since 2021.
The new Mini keeps the same compact 8.3-inch size, but comes with the new A17 Pro chip, which Apple says delivers a 30% boost in CPU performance and 25% on GPU. It also supports the Apple Pencil Pro, and comes with 128 gigabytes of memory (double the amount in the previous version) in the $499 base model.
That extra computing power is key, since Apple says this iPad is purpose built for Apple Intelligence. Of course, as with the latest iPhones, none of the AI features will actually be available until the next version of iOS ships later this month. —Alexei Oreskovic
NYT to Perplexity: Keep your crawlers to yourself
The New York Times does not, I repeat, does not want AI companies to access or use its content—for anything—without permission.
Apparently AI search engine Perplexity did not get the memo. The Gray Lady sent the startup a “cease and desist” notice to get it to stop accessing articles from the New York Times website, according to the Wall Street Journal.
Perplexity has signed revenue-sharing agreements with several publishers, including Time, Der Spiegel, and, yes, Fortune. The deals let Perplexity access content from the media outlets to and, in turn, the outlets' content can be cited in Perplexity’s AI-generated search results.
But the New York Times has no such agreement with Perplexity and already has a pending lawsuit against OpenAI and Microsoft, claiming that they violated copyright law by training their chatbots on millions of its articles without approval.
Perplexity’s CEO Arvind Srinivas responded to the New York Times’ letter by saying Perplexity has “no interest in being anyone’s antagonist here.” But the Times appears to be ready to rumble. —Sharon Goldman
Correction, Oct. 16: A previous version of this article misstated how Perplexity operates. It creates online summaries and adds citations about where that information comes from. As part of its deals with publishers, Perplexity shares some of the revenue it generates with those outlets.
How much is that AI in the window
One big buyer of artificial intelligence? The U.S. government.
The country’s defense and security forces are stocking up on AI and enlisting hundreds of companies to develop and safety-test new AI algorithms and tools, according to new Fortune analysis.
In the two years since OpenAI released the ChatGPT chatbot, the U.S. Dept. of Defense has awarded roughly $670 million in contracts to nearly 323 companies to work on a range of AI projects—a 20% increase from 2021 and 2022.
The Dept. of Homeland Security awarded another $22 million in contracts to 20 companies doing similar work in 2022 and 2023, more than triple what it spent in the prior period.
(The AI companies working with the military include well-known contractors like Palantir and younger startups like Scale AI.)
The military has always supported the development of cutting-edge technology—ARPANET, anyone?—but naturally, the spending aligns with the private sector’s own AI obsession. Giddyup. —AN
Don’t call it ‘resting and vesting’
The Japanese game publisher Bandai Namco is definitely not trying to encourage hundreds of its employees to quit.
The Tekken and Soulcaliber maker, which recently cut some titles, reportedly moved 200 of its 1,300-ish workers into rooms where they have nothing to do—all the way back in April.
As Bloomberg explains, this appears to be a “traditionally Japanese approach” to getting rid of staff without clashing with strict labor laws. People sent to “expulsion rooms” are usually told poor performance—the only outcome, with no tasks—could lead to a reduction in their eventual severance.
But Bandai Namco denies that this is its tactic. “Some employees may need to wait a certain amount of time before they are assigned their next project,” it said.
Apparently nearly half of the workers have taken the not-a-hint and quit. —David Meyer
Everyone should have a role in AI boom
If you’ve heard of Fei Fei Li—and readers of this newsletter surely have—it very likely came with the moniker “Godmother of AI” attached. She bristles at the term, she shared with a packed room Tuesday night at Fortune’s Most Powerful Women Summit in Southern California.
"It really is not my personality to call myself godmother of anything,” Li said. She initially wanted to reject the title entirely—but the history of science and technology is packed with men who are founding fathers of this and godfathers of that. “If women are readily rejecting that title, where's our voice?" she asked.
Li said her 2023 memoir, The Worlds I See, was her attempt to “tell the AI story”—and do it in a way that encourages young women, immigrants, and people from all walks of life to “find a voice in AI.”
Today’s AI conversation “is so hyperbolic,” Li said—global scale, infinite productivity, billions of whatever—and it’s important that people of all backgrounds have a piece of that so “everyone who cares has a role.” After all, she added: “Whoever makes AI changes the world.” —AN
More data
—Android 15 comes to Pixel devices. An app “private space” and theft detection lock are on tap.
—Spotify rolls out music videos in 85 new markets. I want my STV?
—Tech consultant on trial for Cash App founder's death. Nima Momeni pleaded not guilty.
—There's a talent shortage in chips. A million skilled workers are needed by 2030, Deloitte estimates.
—People are using AI “nudify” bots on Telegram. More nonconsensual deepfakes to crack down on.