> It’s a flippant move by Hegseth.

Care to convert this into a prediction?: are you predicting Hegseth will back down?

> I doubt anyone at the Pentagon is pushing for this.

... what does this mean to you? What comes next? As SecDef/SecWar, Hegseth is the head of the Pentagon. He's pushing for this. Something like 2+ million people are under his authority. Do you think they will push back? Stonewall?

One can view Hegseth as unqualified, even a walking publicity stunt while also taking his power seriously.

> are you predicting Hegseth will back down?

I think he may be able to cancel Anthropic’s contract. But no more. He won’t back down as much as be overruled.

> As SecDef/SecWar, Hegseth is the head of the Pentagon

On paper. Also, being the de jure head of something doesn’t automatically mean you speak for it as a whole.

> while also taking his power seriously

Authority and power are different. A plane pilot has a lot of authority. They don’t have a lot of power.

> I think he may be able to cancel Anthropic’s contract.

This outcome might be a win for everyone involved, the time and effort for those billions with a lot of strings attached are less useful as Ai matures.

The above is fairly surface level. See my other comment for particulars that matter a lot: https://news.ycombinator.com/item?id=47176361

You’ll notice I’m trying to avoid debating generic phrases and terms such as “power” that probably won’t advance mutual understanding of this situation. I’m talking about specific actions and systems. It makes it clearer.

> notice I’m trying to avoid debating generic phrases

You’re missing the forest for the trees. Take the tariffs as analogy. Specifying the laws invoked to effect the tariffs is more precise, but less complete than describing Trump, Bessent and Navarro’s motivations and theories.

Same here. We can wax lyrical about the DPA and specific statutory authorities and how they may be litigated. Or we can look at the actual power structures. The former is precise but inaccurate. The latter is the actual dynamic.

> terms such as “power” that probably won’t advance mutual understanding

If terms like power and influence don’t make sense to someone, they’re going to be lost in any political discussion. But particularly under this administration.

There aren’t legal analytic fundamentals driving why Trump hates windmills or Biden pardoned his son, these were expressions of Presidential power and preference. The legality was ex post facto.

Person to person, we’re talking past each other. If we were sitting down face-to-face or even with a video call, this would be a totally different conversation.

How much are we connecting in this particular conversation? What if each of us were to step back and ask 3 questions: What am I trying to communicate? Are we both interested in having this conversation? Are we both learning from it?

Again, this is not meant as a criticism of you. It is a statement of the dynamic here, and how we’re relating. (Even though HN is well above average, it has massive failure modes when you view it from a systems POV.)

My feeling is that you aren’t responding to the intent behind my statement. But I’ll also recognize that I’m probably not communicating that lands for you. Maybe you feel the same in reverse? That would be my guess.

This as a failure of our communication norms and technologies. Given we’re in the year 2026 and have minimal technical barriers, we have very much failed culturally to get anywhere close to the potential of the Internet or whatever needs to come next.

Genuine question, are you using AI to edit your comments? Going on a rhetorical side quest in a straightforward discussion about policy, law and politics is…well, it’s not on topic.

For what it’s worth, I’m not seeing a failure of communication. I’m seeing a failure of scoping. You’re arguing on the basis of specific legal mechanisms by which power is expressed. I’m arguing the real motivations of and political constraints on decision makers are more fundamental in this case.

That isn’t universally true. Power predicted what Trump would do with tariffs (again, analogy). Legal analysis predicted his constraints (which SCOTUS affirmed). In this case, SecDef has the legal authority to do what’s described. He doesn’t, however, have the political freedom to do so. That turns the latter into the germane constraint, not a litany of proscribed powers.

Put another way, the people—here—are fundamental. (Market reactions, too, though again largely because the people in this administration have chosen the Dow as a lighthouse.) The legal justifications are worse than surface level, they’re ex post facto findings of retaliatory paths. It may feel more substantial to quote DPA statute versus discuss Hegseth and Dario’s motivations and relationships, but that’s, again, missing the forest for the trees.

It takes two to tango. I bowed out nicely and put in a good faith effort to communicate why. Maybe on a different day in a different forum, we could have a useful conversation for both of us. I would look forward to that.

[flagged]

Please don't cross into being a jerk. Posts like this one and https://news.ycombinator.com/item?id=47175955 are the kind of thing we ban accounts for, regardless of how right you are or feel you are.

It's true that there's a lot of grey area and turbulence right now around which HN posts have been LLM-generated or LLM-edited, and it's compounded by the fact that there's no way to tell for sure. We all have to find our way through this—both the community and the mods. But we can and need to do so without breaking HN's rules ourselves in the process.

It matters because the whole media is selling this as a Pentagon initiative, while probably 75% in the Pentagon think this is snake oil just like the previous Microsoft VR goggles.

If they don't oppose directly, large bureaucracies know how to drag their feet until the midterms at least, if not until 2028. Soldiers literally dragged their feet at the glorious Trump military parade, when they walked disinterested and casually instead of marching.

> If they don't oppose directly, large bureaucracies know how to drag their feet until the midterms at least, if not until 2028.

While I grant the spirit of this point, I don't think it applies to this situation. The "bureaucratic resistance" explanation doesn't fit when you think about what would happen next. Here is my educated guess based on some research:

- contract termination: Hegseth can direct the relevant contracting officer(s) at the Pentagon to terminate the contract. This could happen within days. Internal stonewalling here might add weeks of delay, but probably not more than that.

- supply chain risk designation: Hegseth signs a document, puts it into motion. Then it becomes a bureaucratic process that chugs along. Noncompliant contracting officers probably would be fired, so this happens within weeks or a few months. Substantial delays could come from litigation, to be sure -- but this isn't a case where civil service stonewalling saves us.

- Defense Production Act: would require an executive order from Trump. This would go into effect right away, at least on paper. It would very likely lead to litigation and possibly court injunctions.

My point is that non-compliant civil servants at the Pentagon probably can't slow it down very much. (I recommend they do what their oath and conscience demands, to be sure!) Hegseth has shown he's willing to fire quickly and aggressively. I admire people who take a stand against Hegseth and Trump -- they are a nasty combination of dangerous and corrupt. At the moment, they appear weaker than ever. Sustained civil pushback is working.

Let's "roll this up" back to my original point. I responded to a comment that said "I doubt anyone at the Pentagon is pushing for this.", asking the commenter to explain. I don't think that comment promotes a better understanding of the situation. It is more useful to talk about the components of the situation and some possible cause-effect relationships.

[flagged]