Thank you for this post. We need to think deeply about what happens when the tools of national power are turned on domestic organizations and companies - export controls (e.g., Anthropic), nationalization (e.g., Intel, MP Materials), non-kinetic obliteration (e.g., USAID and most federal agencies). The research and practical questions are about the strategy (other than tributes) they will use to survive and thrive, and whether they can do so ethically and sustainably through successive regimes. I think you are completely accurate on Dean Ball and whether he really thinks through the consequences of a capricious and ill-informed unitary executive.
Henry Farrell’s article on the Trump-Anthropic/Fable dispute is useful because it shows that the fight over AI regulation is not simply a fight between regulation and deregulation. It is also a fight over *how* regulation is exercised and who benefits from that exercise. Often, the explanation we are given is only the surface story. A policy, decision, or public argument may be presented as national security, public safety, technological progress, market efficiency, or institutional reform. Sometimes that explanation is sincere. But often it is incomplete. Beneath the surface story sits a deeper truth: someone is trying to gain, protect, or redirect power. The real question, then, is not only *what are they saying?* It is also *what benefit does this action produces, who receives that benefit, and who loses power as a result?*
This is where three ideas fit together. **Surface story/deeper truth** helps us see the gap between the stated rationale and the underlying power move. **Distract-to-extract** describes the strategy: create noise, conflict, spectacle, complexity, or crisis so attention is pulled toward the visible drama while power, money, authority, access, or institutional control is extracted elsewhere. **Weaponization of ambiguity** explains the method: use vague laws, unclear procedures, technical complexity, emergency powers, discretionary authority, or legal gray zones to make the extraction difficult to see, challenge, or stop. Farrell’s article matters because it shows this mechanism in action: AI safety and national security may be the surface story, but the deeper issue is the expanding ability of executive power to use ambiguity itself as leverage over strategically important private actors.
Henry Farrell is an illustration that you can like SF (like Vinge's seminal _A Fire Upon The Deep_" and "_A Deepness In The Sky__") and not become a simplistic fool (such as much of the techbro class seems to be with very simplistic beliefs, including researchers such as Hinton, or Sutskever's, illustrated by the latter's idea from a while back that to get Artificial Superintelligence, you could ask an LLM and get that level of an answer by saying "Answer like you are a superintelligence. ..." — yes he really said that in a BBC interview).
I think every C-level executive (markets), every national politician (publics/societies), and every high-level bureaucrat (bureaucracies) should read what Henry Farrell writes, and listen to what he has to say.
Per the conclusion: “We need to think more seriously about the problems that this will create, and indeed is already creating. I’m not seeing much of that thinking yet.” The broader implication of the AI debate is that it isn’t in formal public policy institutions precisely because the Trump administration eschews democratic practices. Lots of us can cringe at the formal regulatory process but it is at least pro forma public. Congressional legislation, ditto. Executive orders and their equivalent in-house decisions, not public. It’s a lousy process and dysfunction which might be the key implication.
Good heavens. I mean, aside from the ever-more-clear fraud aspects of this whole 'enterprise' to bring us the AI Future, it leaves me striving, as an un-noteworthy individual, to be less and less 'connected', and more and more of a 'raw savage' (in Kingsnorthian "Against the Machine" terms), refusing to comply with the exhausting thrust and vibe of the owners of this "New Industrial Revolution" madness whose money for advertising 'just how wonderful they are' exceeds all reason. Oh, and also, their campaign contributions. I'm leaving my dad's old Klein sidecutters hanging by the coax coming off the alley...
I suppose it goes without saying that we're in deep trouble. This regime is doing its damnedest to clamp down on dissent. We need to respond by pushing back even harder. Whatever it takes.
I liked your post in general and your connection between Biden and Trump's approaches to national security; however, I think you have misinterpreted Dean's post. I know him not at all, but I think he is making a statement about how things *are* rather than how they ought to be; he's not pleased by the Trump administration's concentration and abuses of power, but he views it as the status quo and is evaluating Anthropic's actions within that status quo, wherein they did not make much business sense.
Thank you for this post. We need to think deeply about what happens when the tools of national power are turned on domestic organizations and companies - export controls (e.g., Anthropic), nationalization (e.g., Intel, MP Materials), non-kinetic obliteration (e.g., USAID and most federal agencies). The research and practical questions are about the strategy (other than tributes) they will use to survive and thrive, and whether they can do so ethically and sustainably through successive regimes. I think you are completely accurate on Dean Ball and whether he really thinks through the consequences of a capricious and ill-informed unitary executive.
Henry Farrell’s article on the Trump-Anthropic/Fable dispute is useful because it shows that the fight over AI regulation is not simply a fight between regulation and deregulation. It is also a fight over *how* regulation is exercised and who benefits from that exercise. Often, the explanation we are given is only the surface story. A policy, decision, or public argument may be presented as national security, public safety, technological progress, market efficiency, or institutional reform. Sometimes that explanation is sincere. But often it is incomplete. Beneath the surface story sits a deeper truth: someone is trying to gain, protect, or redirect power. The real question, then, is not only *what are they saying?* It is also *what benefit does this action produces, who receives that benefit, and who loses power as a result?*
This is where three ideas fit together. **Surface story/deeper truth** helps us see the gap between the stated rationale and the underlying power move. **Distract-to-extract** describes the strategy: create noise, conflict, spectacle, complexity, or crisis so attention is pulled toward the visible drama while power, money, authority, access, or institutional control is extracted elsewhere. **Weaponization of ambiguity** explains the method: use vague laws, unclear procedures, technical complexity, emergency powers, discretionary authority, or legal gray zones to make the extraction difficult to see, challenge, or stop. Farrell’s article matters because it shows this mechanism in action: AI safety and national security may be the surface story, but the deeper issue is the expanding ability of executive power to use ambiguity itself as leverage over strategically important private actors.
Project 2025 in a nutshell.
Henry Farrell is an illustration that you can like SF (like Vinge's seminal _A Fire Upon The Deep_" and "_A Deepness In The Sky__") and not become a simplistic fool (such as much of the techbro class seems to be with very simplistic beliefs, including researchers such as Hinton, or Sutskever's, illustrated by the latter's idea from a while back that to get Artificial Superintelligence, you could ask an LLM and get that level of an answer by saying "Answer like you are a superintelligence. ..." — yes he really said that in a BBC interview).
I think every C-level executive (markets), every national politician (publics/societies), and every high-level bureaucrat (bureaucracies) should read what Henry Farrell writes, and listen to what he has to say.
Per the conclusion: “We need to think more seriously about the problems that this will create, and indeed is already creating. I’m not seeing much of that thinking yet.” The broader implication of the AI debate is that it isn’t in formal public policy institutions precisely because the Trump administration eschews democratic practices. Lots of us can cringe at the formal regulatory process but it is at least pro forma public. Congressional legislation, ditto. Executive orders and their equivalent in-house decisions, not public. It’s a lousy process and dysfunction which might be the key implication.
Good heavens. I mean, aside from the ever-more-clear fraud aspects of this whole 'enterprise' to bring us the AI Future, it leaves me striving, as an un-noteworthy individual, to be less and less 'connected', and more and more of a 'raw savage' (in Kingsnorthian "Against the Machine" terms), refusing to comply with the exhausting thrust and vibe of the owners of this "New Industrial Revolution" madness whose money for advertising 'just how wonderful they are' exceeds all reason. Oh, and also, their campaign contributions. I'm leaving my dad's old Klein sidecutters hanging by the coax coming off the alley...
Tim Long, Just Up the Hill from Lock 15
I suppose it goes without saying that we're in deep trouble. This regime is doing its damnedest to clamp down on dissent. We need to respond by pushing back even harder. Whatever it takes.
I liked your post in general and your connection between Biden and Trump's approaches to national security; however, I think you have misinterpreted Dean's post. I know him not at all, but I think he is making a statement about how things *are* rather than how they ought to be; he's not pleased by the Trump administration's concentration and abuses of power, but he views it as the status quo and is evaluating Anthropic's actions within that status quo, wherein they did not make much business sense.