The Trump administration just told a company they needed pull their new product. More government interference with private companies? A rational concern and appropriate national security response?
Will the tech bros rebel? I know people that were actively developing with Fable and experienced the shutdown in real time.
Are you in favor of the government telling a company to pull a product? Should AI models have to go through some kind of government sanity certification before release? Would you feel different about it if your alarm didn’t go off tomorrow because an American company’s AI enabled a foreign nation to shut down the cellular networks and power grid?
They didn’t tell them to pull it, they told them to revoke access from non-US nationals. Similar to export control rules for cryptography that have existed for a long time. PGP got its author a criminal investigation back in the day.
Now Anthropic can’t guarantee a restriction to US nationals so Anthropic pulled it instead.
They probably have some internal problems with this too. They just hired Andrej Karpathy (one of the founders of OpenAI) - he’s a Greencard holder but not a US national.
You can be mad anyway of course, but maybe more in the context of a government stake in OpenAI being discussed recently and the conflict of interest this brings. Export control by itself isn’t new though.
And, as the post points out, it was an impossible demand. So it’s functionally the same.
And it’s not the same as export control. I’m old enough to remember encryption export control. There was no prohibition on non-citizens in US territory from using any encryption level they wanted for domestic use. Half the emails and Usenet posts I got back then were by non-citizens residing in the US and using PGP with RSA above the limit.
Export control as a concept isn’t new, but this a very new application of it. The ban seems to apply to all non-US citizens for a whole product version, not only one aspect of that product, or potentially even other versions of the same product.
The ban is also unlike cryptography in that this wasn’t a military tech with regulatory controls that began to be adopted by companies who knew there were restrictions. This is the reverse, a ban applied to a pre-existing application where there were notoriously no controls in place.
Anthropic has multiple offices internationally, so even if we put Karpathy aside for a moment, it would be an unbelievably complex task to silo off internal access and work allocation. It’s not like it’s a Netscape situation where they have to make sure the few people working with the RSA algorithm details are US citizens, and they have to ship a slightly nerfed fork alongside the US version, for a product that works entirely on a local single machine.
I’m certainly not angered by the situation, but it’s very clearly market manipulation with nonsensical justification, and far wider implications than just for Anthropic.
The techbros are one of the major driving forces behind the coup currently capturing the US government. Rebel against who exactly? Trump is in office because of them. He works for them. JD Vance is one of them.
Should AI models have to go through some kind of government sanity certification before release?
AI needs to be heavily regulated, full stop. Its one of the most dangerous pieces of technology that has ever been pushed willy-nilly on humanity.
Would you feel different about it if your alarm didn’t go off tomorrow because an American company’s AI enabled a foreign nation to shut down the cellular networks and power grid?
Why the fuck are you using an alarm that relies on the cellular network?? And if the power grid goes down, there’s much bigger problems than my alarm not going off.
The techbros are one of the major driving forces behind the coup currently capturing the US government.
Thus the paradox of the government shutting down this release.
I think you entirely missed the point about the power going out. Yet, still, your reply proved exactly what I was getting at. This is much bigger than both the unregulated release of AI and the government putting its foot down on the throat of those releasing these threats.
The whole vibe of your comment feels like it missed what I was saying.
The Trump administration just told a company they needed pull their new product. More government interference with private companies? A rational concern and appropriate national security response?
Will the tech bros rebel? I know people that were actively developing with Fable and experienced the shutdown in real time.
Are you in favor of the government telling a company to pull a product? Should AI models have to go through some kind of government sanity certification before release? Would you feel different about it if your alarm didn’t go off tomorrow because an American company’s AI enabled a foreign nation to shut down the cellular networks and power grid?
They didn’t tell them to pull it, they told them to revoke access from non-US nationals. Similar to export control rules for cryptography that have existed for a long time. PGP got its author a criminal investigation back in the day.
Now Anthropic can’t guarantee a restriction to US nationals so Anthropic pulled it instead.
They probably have some internal problems with this too. They just hired Andrej Karpathy (one of the founders of OpenAI) - he’s a Greencard holder but not a US national.
You can be mad anyway of course, but maybe more in the context of a government stake in OpenAI being discussed recently and the conflict of interest this brings. Export control by itself isn’t new though.
And, as the post points out, it was an impossible demand. So it’s functionally the same.
And it’s not the same as export control. I’m old enough to remember encryption export control. There was no prohibition on non-citizens in US territory from using any encryption level they wanted for domestic use. Half the emails and Usenet posts I got back then were by non-citizens residing in the US and using PGP with RSA above the limit.
Export control as a concept isn’t new, but this a very new application of it. The ban seems to apply to all non-US citizens for a whole product version, not only one aspect of that product, or potentially even other versions of the same product.
The ban is also unlike cryptography in that this wasn’t a military tech with regulatory controls that began to be adopted by companies who knew there were restrictions. This is the reverse, a ban applied to a pre-existing application where there were notoriously no controls in place.
Anthropic has multiple offices internationally, so even if we put Karpathy aside for a moment, it would be an unbelievably complex task to silo off internal access and work allocation. It’s not like it’s a Netscape situation where they have to make sure the few people working with the RSA algorithm details are US citizens, and they have to ship a slightly nerfed fork alongside the US version, for a product that works entirely on a local single machine.
I’m certainly not angered by the situation, but it’s very clearly market manipulation with nonsensical justification, and far wider implications than just for Anthropic.
The techbros are one of the major driving forces behind the coup currently capturing the US government. Rebel against who exactly? Trump is in office because of them. He works for them. JD Vance is one of them.
AI needs to be heavily regulated, full stop. Its one of the most dangerous pieces of technology that has ever been pushed willy-nilly on humanity.
Why the fuck are you using an alarm that relies on the cellular network?? And if the power grid goes down, there’s much bigger problems than my alarm not going off.
Thus the paradox of the government shutting down this release.
I think you entirely missed the point about the power going out. Yet, still, your reply proved exactly what I was getting at. This is much bigger than both the unregulated release of AI and the government putting its foot down on the throat of those releasing these threats.
The whole vibe of your comment feels like it missed what I was saying.