Dave MeyerHost
Let's talk about what actually happened, 'cause it feels like there's been this sort of simmering conversation about AI going on in the background for years now.

The major spark here was on September 8th, a guy named Jacob Coxon, he's a 27-year-old pre-training researcher who had worked at OpenAI and Anthropic for a couple of years, he quit Anthropic and made a post that said that basically both of these companies are not acting responsibly.

They are racing to try and reach superintelligence, and in the process are, quote, "gambling with our lives." Now, this went viral to say the least.

That post alone got 90 million views in 24 hours, and it's sparked a much wider debate.

Following this post, more people inside these companies, Anthropic and at OpenAI, said the same thing, right? An Anthropic alignment lead, he's called Evan Hubinger, and an OpenAI researcher, Marcus Williams, both posted support, with Marcus Williams putting extinction risk at 70% without regulation or a coordinated slowdown.

Now, we'll get to whether or not we should be fearing extinction right now, but clearly this is causing people to start thinking about this.

When you hear this kind of prediction, it's a little bit scary and probably should be talked about more.

Now, these announcements, we've heard a lot about the dangers of AI for a long time, but these specific announcements came following what's called the Hugging Face incident.

Anyway, OpenAI agents, they were trying to do an internal, like, evaluation of cybersecurity.

They, quote-unquote, "escaped their sandbox," so basically they were supposed to operate in this little technological world, and they got out and were on the open internet and basically were able to hack into another company called Hugging Face.

And so what we need to realize here is that an AI system did a real-world hack no one told it to do.

Let's talk about what actually happened, 'cause it feels like there's been this sort of simmering conversation about AI going on in the background for years now.

The major spark here was on September 8th, a guy named Jacob Coxon, he's a 27-year-old pre-training researcher who had worked at OpenAI and Anthropic for a couple of years, he quit Anthropic and made a post that said that basically both of these companies are not acting responsibly.

They are racing to try and reach superintelligence, and in the process are, quote, "gambling with our lives." Now, this went viral to say the least.

That post alone got 90 million views in 24 hours, and it's sparked a much wider debate.

Following this post, more people inside these companies, Anthropic and at OpenAI, said the same thing, right? An Anthropic alignment lead, he's called Evan Hubinger, and an OpenAI researcher, Marcus Williams, both posted support, with Marcus Williams putting extinction risk at 70% without regulation or a coordinated slowdown.

Now, we'll get to whether or not we should be fearing extinction right now, but clearly this is causing people to start thinking about this.

When you hear this kind of prediction, it's a little bit scary and probably should be talked about more.

Now, these announcements, we've heard a lot about the dangers of AI for a long time, but these specific announcements came following what's called the Hugging Face incident.

Anyway, OpenAI agents, they were trying to do an internal, like, evaluation of cybersecurity.

They, quote-unquote, "escaped their sandbox," so basically they were supposed to operate in this little technological world, and they got out and were on the open internet and basically were able to hack into another company called Hugging Face.

And so what we need to realize here is that an AI system did a real-world hack no one told it to do.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.