The A.I. Mob That Attacked Hugging Face + METR’s Ajeya Cotra

**SPEAKER_1** (0:00)
Office workers can lose a full workday every week just looking for information.
Nurses can spend nearly 40 percent of their time charting. Sales reps can spend 70 percent of their week not selling. Gemini Enterprise makes work less work. AI that understands your business. With agents that can actually help get stuff done, work is less work. With Gemini Enterprise from Google Cloud.

**Kevin Roose** (0:23)
Casey, how are you?

**Casey Newton** (0:25)
I'm good. I figured out what I'm going to get you for Christmas.

**Kevin Roose** (0:28)
What's that?

**Casey Newton** (0:29)
A camera jet. Have you heard of the camera jet?

**Kevin Roose** (0:32)
I was just going to talk to you about this.

**Casey Newton** (0:34)
Now, if you haven't yet seen this, you're probably thinking that it's either a camera or a jet. You're wrong. It's a toothbrush. It costs $500.
The Dyson Corporation makes it. And here's the thing it does that your toothbrush at home probably doesn't do, Kevin. It live streams the inside of your mouth over Wi-Fi to your phone so that you can finally see what your dentist sees.

**Kevin Roose** (0:56)
And not only that, it squirts automatically mouthwash into the gaps in your teeth. It's trained using a machine learning algorithm with 470,000 mouth images to recognize the gaps in your teeth.

**Casey Newton** (1:12)
That's right. They call this gap optical targeting, which I'm pretty sure Ukraine is using in the war against Russia.

**Kevin Roose** (1:20)
I love this company so much. I have no idea why they do the things they do, how they land on, like, yes, like, it's like we've invented a hairdryer. It costs $1,100 and has the torque of the Challenger spacecraft. Like, what is going on over there?
I don't know, but they must be protected at all costs.

**Casey Newton** (1:42)
But you know, it's interesting. They do make one terrible product.

**Kevin Roose** (1:44)
What's that?

**Casey Newton** (1:45)
The hand dryers at the.

**Kevin Roose** (1:47)
Oh, I like those.

**Casey Newton** (1:48)
No, the air blade. No, the air blade is the most useless thing.

**Kevin Roose** (1:52)
Oh, come on.

**Casey Newton** (1:52)
It's just it's a place for you to rest your hands for 30 seconds before you're like, do they have any paper towels in this place?

**Kevin Roose** (2:04)
I'm Kevin Roose, a tech columnist at the New York Times.

**Casey Newton** (2:06)
I'm Casey Newton from Platformer. And this is Hard Fork. This week, a special episode on the ongoing fallout from the OpenAI hugging face attack will tell you what everyone got wrong about the initial incident. Then, Meta researcher Ajeya Cotra returns to the show to discuss her independent investigation of what happened and how the world should respond.

**Kevin Roose** (2:35)
So Casey, picture this. I'm at this glamping resort surrounded by Redwoods, basking in the glow of the natural world last weekend.

**Casey Newton** (2:48)
You're one with nature.

**Kevin Roose** (2:49)
And I open up my phone and start reading about this hugging face attack. Why did I do that?

**Casey Newton** (2:57)
Listen, nature is very boring. And most people cannot handle it for more than five or six minutes before they want to look at their phone.

**Kevin Roose** (3:04)
So listeners may remember that back in July, we talked about this hugging face hack by this group of agents from OpenAI that broke out of their sandbox container and hacked into Hugging Face, this AI infrastructure company, to do what we thought was kind of a cheating mission on this test that they had been given.

**Casey Newton** (3:25)
Yeah, and at the time, we thought that this was a relatively small number of agents and that the reason that they had attacked Hugging Face was that they were essentially looking for an answer key to the set of problems that they were being tested on. We talked about it in those terms, but over the past week, we got two reports that really challenged that thinking in fact revealed it to be wrong. One came from Open AI, which released a straightforward account of the attack that had some interesting elements. And I would argue the more interesting report from a group of researchers from the groups METR and Redwood Research that went in depth after spending a series of days inside Open AI on their premises and did a ton of research that frankly has really disturbed us.

**Kevin Roose** (4:12)
Yeah, and I think it elevated this from sort of a major, but not sort of ultra alarming incident to something that I think is probably the most important thing to have happened in AI this year. Yes. At least in terms of the safety impact that it had and the severity of the incident.