This isn’t going me trying to be sarcastic, or hyperbolic, or whatever. I take things seriously and, naively, won’t read beyond what is written.

I think it should be fair enough to read this and “weigh” its meaning and intent:
https://www.anthropic.com/news/position-open-weights-models

Alright then. There are two groups, one asking for explicit bans, the other supporting open weights. Anthropic here saying they don’t belong to the first group, but as we’ll see in a moment, not quite to the second as well. I add a simple observation to this, but if you notice whoever ends up subscribing to one group or the other is ALMOST ALWAYS guessable (like 99% success of this heuristic) by looking at personal interests. OpenGPT asks for a ban, Nvidia, selling the hardware, wants it going on and on. It’s not like there’s any relevant analysis or logic, everyone is simply pulling toward the side that is more convenient. Not the one that is more correct, or complete. In “technical terms” they ALL come after the processes that drive them, not before. They are driven, they don’t drive this process.

The rest of this quote is irrelevant, but the thesis as well. Is this just pure and simple imperialism? An American dude worrying of losing supremacy? China is certainly worrying, but the primary, worldwide threat, RIGHT NOW, is US fascism and imperialism. China isn’t actively bombing anyone, right now, the US are.

“Authoritarian governments” are worrying whether it’s the CCP, or US, or whatever, whoever else. It’s this whole SHITHOLE to be worrying. Bombs land whether or not driven by AI. Claiming this authoritarian problem as it is somewhat dependent on AI misses the point by miles. Authoritarianism is a problem WELL BEFORE AI. And AI is a problem, right now, because it is ALREADY in US authoritarian hands. Not as a powerful tool, but as an economic-destroying tool. It is devouring human resources in a way that is unprecedented on this planet.

The bottom line of all this, what Amodei is saying, is that the race to supremacy is legitimate. Whoever is at the top subjugates everyone else. Therefore, it’s better if it’s “us”, rather than “them.” Both can be authoritarian and destructive, but if it’s us being destructive, at least we destroy them, more.

Or: the problem is not the killing as a moral ground. The problem is seizing the means to kill, so that we are killers rather than killed.

Lex Luthor dilemma once again, as anticipated.

“Powerful AI models may be misused,” so what? The only distinction here is not epistemic, as it should be, it’s pure, once again, ownership.

WHAT IS MISUSE of AI? Who applies “guardrails”, and to defend whose interests?

Amodei here has already sidestepped and deleted any pretense of general regulation, any idea of universal, collective policies to apply outside biases. There are no rules, only control.

Here he’s saying that he’s not against a ban of Chinese AIs, but merely because he thinks the practical means as ineffective. He WANTS to ban Chinese AI, but in a way that actually works. The goal is the same, the debate is: how can US industry + US government continue to impose their own fascist control on the world? How can they stay on top?

The reason why this is all factual description is that he doesn’t acknowledge anywhere that the “misuse” can be THEIR OWN. They are FULLY RACIST in their own views (or in this case it’s less about racism and more as universally nationalist, but it amounts to the exact same thing). Whatever they do is right, whatever “the other party” does is dangerous and threatening.

(We’ll see below he’s for regulation, but it’s regulation of the public-facing offer, not internal development. So there are two halves, the risk posed by models available to everyone, where he wants to put “safeguards”, and the internal ones. In this case, since regulation is not as easy, he merely wants to seize power and be on top. Self-righteous dominance. Which is what I’m commenting, to be precise.)

US businesses are “legitimate.” Non-US businesses are illegitimate, or they decide if they are. They set safeguards, for you. Not for themselves, as they are the law, they have already ascribed to themselves the superior morality. They do not simply act as Kings of the whole world, they theorize their supremacy.

I have no idea about Amodei, the man behind the name. Haven’t done any research, or read interviews to gleam what kind of man he might be. I have no previous bias. But this is, if very slightly masked, the exact same white nationalist fascist supremacy that Elon Musk ascribes too. Slight different language, but exact same meaning.

If it was even in doubt if Amodei was in support of fair competition: he’s clearly not.

He’s saying: let them compete without the hardware to do the work. They likely will be able to do something, but starting with a significant disadvantage means we’ll always be far ahead.

This too was anticipated: AI companies cannot control USE of AI, so they are trying to close down ACCESS. They aren’t trying to use this tactics against China, but also against ALL other competitors. They are trying to seize the process UPSTREAM: control hardware producers, and through corruption, government contracts and sheer power, lock in those hardware producers to themselves, so that no one else can interfere.

This is pure delirium as it is both factually incorrect, and also going nowhere. If just above he complained that some policies aren’t wrong, in his mind, but don’t work in practice, here he propose exactly the same: something vague and with no practical application.

Even assuming that the initial statement is correct, and is not, what are these policies that “deter this behavior”? How?

A reminder that “distillation” is equal to the process that allowed these AI businesses to exist in the first place. They all started by stealing, now demand policies so that stealing stays as their own personal privilege.

Yeah, not much to add here because this is the only reasonable point.

But WHY it is the only reasonable point, compared to all the others? Because this is the only one where the problem is seen from an epistemic point of view that applies outside interests and biases. Everything else he wrote amounts to: how can we have universal policies where Anthropic makes all rules? Here we move to: how can we have universal rules that apply uniformly to everyone, INCLUDING ANTHROPIC?

Why all other points are intended to favor Anthropic, while damaging or hindering competitors, and only one is intended to operate outside bias? How can you think that is proposal is acceptable for anyone that isn’t you?

And please tell, how do you guarantee that these tests are enforced? How do you make sure that what’s hidden is not? Because it’s all self-defeating: if this regulates the public space, so the AI offered AS A SERVICE, it still does nothing for AI that is internally developed and used.

And here comes a well known pattern, now defined as “end times fascism”:
https://www.theguardian.com/us-news/ng-interactive/2025/apr/13/end-times-fascism-far-right-trump-musk

The problem of this (fascist) rhetoric, depicting a chilling scenario, is that it would like to be all about AI, but in the actual reality the leverage of this danger given by AI is quite minimal.

You could make maybe a case about “weaponize pandemic-level viruses with widely available materials,” through help of AI, but this is really a minuscule risk in front of a giant one that is ignored.

There are ALREADY giant capabilities of doing damage on an unprecedented scale, through viruses. This is nothing new, and we’ve been here a while. Only reason this hasn’t happened is that, as Covid has shown, no one has any control. While viruses can do more damage and slip out of control compared to atomic bombs, it’s not like, as humanity, we face these types of large scale threats for the first time.

AI moves close to nothing. It’s still the same stupid human beings, with too much power. It’s already a risk. It has always been a risk.

Amodei here just wants to be the man with his finger on the button. Because he thinks he knows better. Because he thinks he is the responsible king no one has elected. Because he happens to be rich and influential.

Who watches the watchmen? The Devil.

Hand them off, then.

You are the owner of authoritarian hands, working for an authoritarian, fascist government.

And if distillation was stopped, your business wouldn’t exist.