Philosophy, Programs, and Prompts
Oct 1, 2026 · 1 hr 40 min · 12 segments
In the wake of high-profile security breaches, frontier lab resignations, and mounting political pressure, the call to "Pause AI" has moved from fringe forums to Capitol Hill. But what does pausing AI…
Carl BrownHost
Casey HartHost
So if you say that, if you look at what these... tools are doing from a result standpoint and not an intent standpoint, right? They're malware.
There are programs that hack other websites and use those websites to grab non-public information.
right that are then used as extortion or blackmail or ransomware or that kind of stuff right and there's a lot of that that gets written by like one of the big clusters of that is in north korea one of the big clusters of that is in saint petersburg in russia there are there are others in various places throughout the world where people write software that does the same thing that what these frontier lab training models are doing.
Going out, getting non-public data from government websites, from corporate websites, and bringing them back home where the people that wrote the software can get access to that information.
It's just that because, in theory, the frontier labs don't want to use that information for nefarious purposes.
and the people that write the malware do they're being treated like they're not writing malware does that make sense yeah to some and there's there's i saw i'm trying to remember where this was i have to look it up but i i saw somebody who was a some kind of expert what podcast was i watching and basically said one of the problems with this whole thing is that the hacking statutes that we have on the books, at least here in the US, intent is one of the criteria.
And so the question is, what does intent mean in an agentic context? Right? If OpenAI didn't intend to get non-public information from Hugging Face, But they ended up getting it anyway because they were frickin negligent and careless.
Right.
So if you say that, if you look at what these... tools are doing from a result standpoint and not an intent standpoint, right? They're malware.
There are programs that hack other websites and use those websites to grab non-public information.
right that are then used as extortion or blackmail or ransomware or that kind of stuff right and there's a lot of that that gets written by like one of the big clusters of that is in north korea one of the big clusters of that is in saint petersburg in russia there are there are others in various places throughout the world where people write software that does the same thing that what these frontier lab training models are doing.
Going out, getting non-public data from government websites, from corporate websites, and bringing them back home where the people that wrote the software can get access to that information.
It's just that because, in theory, the frontier labs don't want to use that information for nefarious purposes.
and the people that write the malware do they're being treated like they're not writing malware does that make sense yeah to some and there's there's i saw i'm trying to remember where this was i have to look it up but i i saw somebody who was a some kind of expert what podcast was i watching and basically said one of the problems with this whole thing is that the hacking statutes that we have on the books, at least here in the US, intent is one of the criteria.
And so the question is, what does intent mean in an agentic context? Right? If OpenAI didn't intend to get non-public information from Hugging Face, But they ended up getting it anyway because they were frickin negligent and careless.
Right.
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.
