Given open source is lagging frontier labs by 3-5 months, what happens when there’s an accessible model that’s deemed “too dangerous” for public use?
Is the hope that the buffer, enabled by the staggered releases, allows the US gov/corporates to prepare for adversarial use of these models?
Won’t the 3-5 month lag only compress as these open source models stop relying on distillation?
Amid this society-defining race to AGI people still think this isn’t going to be a government project?
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: https://t.co/bwn0sximKZ
How much will 500+bn clusters (Stargate, Colossus) spend on security infra?
There’s no question these clusters will be the foundation on which we achieve AGI (if we’re not already there), and the physical security question seems largely unaccounted for.
If @leopoldasch thesis holds, the state will step in, treating these clusters as true national security assets, but it’s far more difficult to do so retroactively.
Is the tradeoff between getting compute online and ensuring that it’s built on secure infra actually zero-sum?
Me, senior year of college: playing beer pong and writing a thesis about medieval Sicily that zero ppl read
My cousin @NelsonBellows, senior year: speaking with @TaylorLorenz about his hackathon-winning new prediction market app for short-form video 🤯
https://t.co/WehVbRMA0p