Artificial Intelligence

The US, UK, and other major powers (notably excluding China) unveiled a 20-page document on Sunday that provides general recommendations for companies developing and/or deploying AI systems, including monitoring for abuse, protecting data from tampering, and vetting software suppliers.

The agreement warns that security shouldn’t be a “secondary consideration” regarding AI development, and instead encourages companies to make the technology “secure by design”.

When SAG-AFTRA finally reached a tentative agreement to bring the actors strike to an end earlier this month, the union’s insistence that its members might not be able to look over the deal before voting on it raised more than a few eyebrows.

It seemed odd that union leadership might try to get people on board with the contract without giving them a chance to read the whole thing. But now the union’s released the full document — on the Friday evening after Thanksgiving of all days.

This Wall Street Journal article about the recent drama at OpenAI contains an amazing anecdote. Apparently an employee at AI rival Anthropic thought it’d be funny to send “thousands of paper clips in the shape of OpenAI’s logo” as a prank, in reference to the infamous paperclip maximizer thought experiment.

Weirdly, I think OpenAI’s logo makes for a great paperclip design. Should we be worried?

Reports from Reuters and The Information Wednesday night detail an OpenAI model called Q* (pronounced Q Star) that was recently demonstrated internally and is capable of solving simple math problems. Doing grade school math may not seem impressive, but the reports note that, according to the researchers involved, it could be a step toward creating artificial general intelligence (AGI).

After the publishing of the Reuters report, which said senior exec Mira Murati told employees that a letter about Q* “precipitated the board’s actions” to fire Sam Altman last week, OpenAI spokesperson Lindsey Held Bolton refuted that notion in a statement shared with The Verge: “Mira told employees what the media reports were about but she did not comment on the accuracy of the information.”

Separately, a person familiar with the matter told The Verge that the board never received a letter about such a breakthrough and that the company’s research progress didn’t play a role in Altman’s sudden firing.

The drama continues!

The company tested baking a cryptographic “digital signature” into photos taken by its cameras to set them apart from AI-generated or otherwise faked images. Sony says the feature will come to cameras like the Alpha 9 III via a firmware update in Spring 2024.

The company’s Q3 2024 earnings reveal $18.12 billion in revenue for the quarter, with $14.51 billion from its data center business as the world’s AI companies jockey for Nvidia’s chips. That overall revenue number is a 206 percent jump from last year and marks a continuing growth trend for the company.

It also noted China and other areas make up 20 – 25 percent of its chip sales, which new US export restrictions will cut into soon; however, the decline should be “more than offset” by growing sales to other regions.

Here’s a brief update on where things stand with OpenAI today, after an explosive weekend and a very confusing Monday which saw the news of Sam Altman going to Microsoft slowly fade into Satya Nadella not seeming so sure that would happen.

— We’re told Altman still wants to return to OpenAI and continues to negotiate with the board today.

— As Bloomberg reported late last night, new interim CEO Emmett Shear is involved in mediating these negotiations, creating the frankly unprecedented situation where (1) the interim CEO who replaced (2) the interim CEO who replaced Sam and who (3) got replaced for trying to get Sam back is now (4) deeply involved in a new effort to get Sam back. Read it through a few times, it’s fine. It doesn’t make any sense to anyone else either.

— Microsoft’s offer to hire everyone who threatened to quit is still on the table, and has now been made officially public, after being noted in the employee walkout letter yesterday. In general, Microsoft appears to have receded from the situation; Nadella remains in the mix but has now made several media appearances reiterating that he’ll will work with Altman and OpenAI “irrespective of configuration,” which frankly sounds like he’s talking about the benefits of plug and play device drivers in Windows. We all fall back to what we know.

— We are told everyone, including the board, is trying to be reasonable, and put OpenAI back together.

We’ll keep posting updates as we have them; at the very least we can say the overall temperature has dropped, but it’s not clear any of this results in an actual return.

Microsoft CTO and EVP of AI Kevin Scott says the software maker will match OpenAI’s compensation to employees that want to join Sam Atlman’s new AI research lab. It comes shortly after Microsoft CEO Satya Nadella went on a media tour last night and didn’t seem to know if Sam Altman will actually join Microsoft or not. Competitors like Salesforce have been trying to tempt OpenAI employees to join rival AI projects.

Several days after Sam Altman was removed from his post as OpenAI’s CEO, there is still no resolution to the issue, even after hundreds of employees signed a pledge to follow him to Microsoft unless he’s reinstated.

Now, Bloomberg reports that global affairs VP Anna Makanju sent a memo to staff confirming that negotiations have continued between Altman, the remaining board members, and new interim CEO Emmett Shear, they won’t have a final response tonight and plan to continue discussions on Tuesday morning.

As hundreds of employees threaten to quit unless Sam Altman’s firing is rolled back:

The Information reports that before interim CEO Emmett Shear took the job, OpenAI’s remaining board members offered it to two others: former GitHub CEO Nat Friedman and Scale AI CEO Alex Wang. Both turned it down. (Update: And in another report, it says board members also pursued the CEO of its competitor Anthropic, Dario Amodei, and proposed a merger between the two companies.)

From Business Insider there’s a report Ilya Sutskever gave employees two explanations for Altman’s firing: that he gave two people the same project and that he allegedly gave two board members different opinions about a member of personnel.

Finally, the Wall Street Journal writes that Sutskever, one of the board members who voted to fire Altman, switched to asking for Altman’s return “after an emotionally charged conversation with Anna Brockman, Greg Brockman’s wife” and notes he officiated their civil ceremony at OpenAI’s offices in 2019.

We tried to just tell the story in order, which is tough to do. Then we tried to figure out how this happened, what it means, and what might happen next. Grab the audio for your preferred podcast app right here.

And because this story won’t stop, we didn’t even get all the news – because one outcome we didn’t think of is that Sam Altman might go back to OpenAI after all. Sure!

No, really: this does not surprise me at all! Unlike 404’s Jason Koebler, I did in fact read early parts of blogger and AI “doomer” Eliezer Yudkowsky’s Harry Potter and the Methods of Rationality — it was known in nerd/fan circles, not just the effective altruist movement, and at least initially was built around the pretty funny conceit of “applying cold hyperlogic to the quirks of a children’s fantasy series.” I did not get nearly far enough to reach the namecheck of Emmett Shear, OpenAI’s new CEO.

Microsoft employees are waking up to the surprise news of former OpenAI CEO Sam Altman joining them as a colleague. Microsoft CEO Satya Nadella says Altman is joining to lead a new advanced AI research team, but he’s not yet showing up in Microsoft’s corporate directory according to multiple sources. He’s still marked as an OpenAI guest. Once he’s fully onboarded Altman will have a CEO title inside Microsoft and could be set to hire hundreds of OpenAI employees who have threatened to resign.

📩 Do you know more about Microsoft’s plans for its new AI research team? You can reach me confidentially on the Signal messaging app: +442081230413, or through email (tom@theverge.com).

OpenAI chief scientist Ilya Sutskever, who reportedly led the push to fire Sam Altman, which kicked off an entire weekend of shenanigans including three CEOs in three days and Altman joining OpenAI financial backer Microsoft, has regrets. In a post on X, the platform formerly known as Twitter, he writes:

I deeply regret my participation in the board’s actions. I never intended to harm OpenAI. I love everything we’ve built together and I will do everything I can to reunite the company.

Moments after Sutskever’s post, newly-minted Microsoft exec Sam Altman quoted the post with a heart.

OpenAI’s workforce is voicing complaints on X (formerly Twitter) following the ousting of former CEO Sam Altman and the confirmation on Sunday that he wouldn’t be re-instated.

“Every single person whose ever told me to trust them on blind faith has been a liar,” said OpenAI technical staff member Brydon Eastman in a now-deleted post. “A chief scientist should [k]now this better than anyone.”

That’s a big deal in itself. Microsoft typically uses the CEO title for the leaders of big divisions like Microsoft Gaming, or acquired companies like LinkedIn and GitHub. Altman as the CEO of a new advanced AI research team signals to me that Microsoft is treating this like a big acquisition, which makes sense when you consider that a number of OpenAI employees are joining Altman at Microsoft.