Meta extends political ad ban beyond Election Day

The cover—a monochrome portrait of Jobs, closed hand to chin, with a multicolor Fortune logo layered over it—has become an instantly recognizable icon of tech. So much so, in fact, that the HBO comedy series Silicon Valley created its own version for Hooli CEO Gavin Belson’s conference room in 2016.

Take a break from the U.S. Election Day madness by reading Adam Lashinsky’s great profile from 2009. The time Jobs almost took Apple private with Larry Ellison, the Founder Mode-before-it-had-a-name behavior—it’s all in there. But please, if you’re eligible: Vote. —Andrew Nusca

P.S. Gita Gopinath, the IMF’s first deputy managing director, will discuss the shifting global economic landscape at Fortune Global Forum, November 11 to 12 in New York City. Interested? Request an invitation here.

Want to send thoughts or suggestions to Data Sheet? Drop a line here.

Meta extends political ad ban beyond Election Day

Meta originally said it would ban new political ads on Facebook and Instagram through Election Day. Now it’s extending the ban “until later this week,” according to a new update posted to its Business Help Center.

The company neither offered a specific date nor a reason why it extended its ban. One guess would be that Meta seeks to curtail any attempt at courting public opinion while votes are still being counted. The increase in early and mail-in voting, plus varying state laws on when counting can begin, means tallying all votes in closely contested states could take several days.

Meta first announced the ban in August. New ads about social issues, elections, or politics would be barred from Oct. 29, and those with existing ads would have “limited editing capabilities.” The company observed a similar restriction during the 2020 U.S. election, reasoning that there wasn’t enough time to properly contest claims made in ads so close to the election. 

Alphabet, Google’s parent company, announced a similar ad policy last month. Election ad spending is a key driver of revenue for both companies this quarter. —AN

OpenAI making moves to become for-profit

It said it would do the thing. Now it’s doing the thing.

OpenAI is reportedly in talks with regulators in California and Delaware to restructure the company into a for-profit benefit corporation. Under scrutiny, according to Bloomberg: OpenAI’s valuation of its “highly lucrative intellectual property,” including the ChatGPT app.

The company is presently a nonprofit organization with a for-profit arm, or as it likes to say, a “capped profit” structure. It’s been this way since 2019. A new structure would remove the control of OpenAI’s non-profit board as well as the cap on investment return to investors, reshaping it to be more like rivals Anthropic and xAI.

All of this is critical given the company’s recent $6.6 billion fundraise that values the company at $157 billion.

OpenAI’s original nonprofit structure was to ensure that it would create “safe AGI that is broadly beneficial,” using the acronym for artificial general intelligence. With revenue in its sights and convertible debt in its coffers, expect OpenAI’s new structure to be optimized for “popular AI that is broadly lucrative.” —AN

Palantir brightens annual outlook for the third time

It’s been no secret that Palantir Technologies has benefited from the AI boom. But raising its annual forecast for the third time? Even a company named for J.R.R. Tolkien’s fictional seeing stones couldn’t see that coming.

The secretive data analytics company cofounded by Peter Thiel and led by Alex Karp says it’s seeing strong spending from governments and rising demand from businesses for all things generative AI. It now expects its 2024 revenue to be about $2.81 billion, up from its previous prediction of about $2.74 billion.

Palantir shares jumped 9%, to about $47, on the news. Stock for the Denver company, which was added to the S&P 500 in September, is up more than 140% year to date. 

There’s little sign of a slowdown. As one research analyst told Reuters, Palantir’s commercial business will likely overtake its government business as corporations step on the gas around AI software. —AN

Meta makes Llama AI available for U.S. military use

Meta will let the U.S. military and defense contractors use its Llama AI models, the licensing terms for which usually forbid military use.

“Meta wants to play its part to support the safety, security and economic prosperity of America—and of its closest allies too,” policy chief Nick Clegg wrote in a blog post.

It’s not just military contractors that get to use Llama, as Meta said it’s partnering with a host of other companies—from Accenture and Deloitte to Microsoft and AWS—to get the model into various government agencies.

Clegg argued that all this was a great demonstration of how “widespread adoption of American open source AI models serves both economic and security interests.”

Which would sound a bit less ironic if Chinese researchers hadn’t recently adapted a Llama model for their own military purposes. —David Meyer

Netflix will remove nearly all interactive specials

Some things are best left on the cutting room floor. 

Netflix will reportedly remove all but four of its “interactive specials” by Dec. 1 as it winds down a format experiment it launched in 2017. Netflix said in January that it would no longer produce new interactive specials.

Netflix debuted the interactive special seven years ago as the motion picture equivalent of a “choose your own adventure” book. Starting with Puss In Book: Trapped In An Epic Tale, viewers could make a series of decisions that would alter the trajectory of the narrative. 

Dozens of titles followed, from the kid-oriented (Carmen Sandiego: To Steal or Not to Steal) to the decidedly adult (Black Mirror: Bandersnatch). Now most of those titles will disappear for good. 

The ones that will remain for now, according to The Verge: Unbreakable Kimmy Schmidt: Kimmy vs. the Reverend, Ranveer vs. Wild with Bear Grylls, You vs. Wild, and Bandersnatch, which pushed the format to dizzying complexity in 2019. (“As artificial as the construction may be, I can’t help but admire, just a little bit, this spin on interactive gaming/movie-watching,” wrote Aisha Harris in the New York Times.)

Lessons from the ill-fated experiment? Expect them to pop up in Netflix’s narrative gaming forays, rather than its conventional movie efforts, one exec told Game File. —AN

More data

—Physical Intelligence raises $400m. Robot AI company is backed by Jeff Bezos, OpenAI, Thrive.

—New York Times software developers strike. Hundreds dig in ahead of Election Day.

—“Infostealer” malware is rampant. Behind breaches at Ticketmaster, AT&T, Santander, EA.

—GM convinced ditching CarPlay, Android Auto is the right thing. It’s about controlling the stack.

—Women see more from Kamala Harris on TikTok. Users who identify as female saw 40% more Harris campaign videos than those identified as men, according to a survey.

Endstop triggered