169
Anthropic says it ‘cannot in good conscience’ allow Pentagon to remove AI checks
(www.theguardian.com)
Why not link the actual company's statement:
this post was submitted on 27 Feb 2026
169 points (100.0% liked)
politics
28581 readers
2689 users here now
Welcome to the discussion of US Politics!
Rules:
- Post only links to articles, Title must fairly describe link contents. If your title differs from the site’s, it should only be to add context or be more descriptive. Do not post entire articles in the body or in the comments.
Links must be to the original source, not an aggregator like Google Amp, MSN, or Yahoo.
Example:
- Articles must be relevant to politics. Links must be to quality and original content. Articles should be worth reading. Clickbait, stub articles, and rehosted or stolen content are not allowed. Check your source for Reliability and Bias here.
- Be civil, No violations of TOS. It’s OK to say the subject of an article is behaving like a (pejorative, pejorative). It’s NOT OK to say another USER is (pejorative). Strong language is fine, just not directed at other members. Engage in good-faith and with respect! This includes accusing another user of being a bot or paid actor. Trolling is uncivil and is grounds for removal and/or a community ban.
- No memes, trolling, or low-effort comments. Reposts, misinformation, off-topic, trolling, or offensive. Similarly, if you see posts along these lines, do not engage. Report them, block them, and live a happier life than they do. We see too many slapfights that boil down to "Mom! He's bugging me!" and "I'm not touching you!" Going forward, slapfights will result in removed comments and temp bans to cool off.
- Vote based on comment quality, not agreement. This community aims to foster discussion; please reward people for putting effort into articulating their viewpoint, even if you disagree with it.
- No hate speech, slurs, celebrating death, advocating violence, or abusive language. This will result in a ban. Usernames containing racist, or inappropriate slurs will be banned without warning
We ask that the users report any comment or post that violate the rules, to use critical thinking when reading, posting or commenting. Users that post off-topic spam, advocate violence, have multiple comments or posts removed, weaponize reports or violate the code of conduct will be banned.
All posts and comments will be reviewed on a case-by-case basis. This means that some content that violates the rules may be allowed, while other content that does not violate the rules may be removed. The moderators retain the right to remove any content and ban users.
That's all the rules!
Civic Links
• Congressional Awards Program
• Library of Congress Legislative Resources
• U.S. House of Representatives
Partnered Communities:
• News
founded 2 years ago
MODERATORS
I am going to buy a big bucket of pop-corn for this
Just give them a fake broken version of it and say "we are still working on some of the bugs".
give them a fake broken version
So normal AI then?
Haters so far behind at this point they don't understand how they're getting absolutely fucked by ai agent swarms.
An individual agent is stupid, a million of them can get a job done that is otherwise impossible.
They're being used for everything from marketing to propaganda at this point, and most people still think they suck because they can be tricked or fail basic questions.
You can trick humans too, doesn't mean they can't do impressive things as well, especially in large groups.
Ah yes, the impossible task of creating a barely functioning C compiler...
The funny thing about fascists is that they are so mentally ill, immature, and authoritarian that they self-own constantly. The traits that make them fascist also render them completely incompetent and unable to create; only steal, co-opt, terrorize, and destroy.
If they nationalize or criminalize Anthropic, they will lose most of the brains that made it competitive, and be left with yet another failed fascist endeavor. It's in their best interests to compel companies with a lighter touch. This is why every bit of resistance matters, from every institution and org (though don't expect late stage caps to do the right thing).
But most Americans still do not understand the only way out of fascism is by physically removing the fascist cancer, so I hope they do try to nationalize Anthropic and continue the last year of accelerationism. Losing everything is the only scenario where Americans will revolt against the fascist dictatorship, and that must happen before they have fully autonomous drone killbots to mass murder all who oppose them.
💯
Would YOU trust Kegsbreath not to use it anyways and then try and shift the blame when it goes haywire? Perhaps you should go watch Robocop (80’s version) to see how that works out?
i give it a week before they 180 on this
I swear there was an article a few days ago saying they would knuckle under.
"Not without additional massive investment into our research department so we can responsibly work on countermeasures."
I'm not finding that part?
Well, that’s scary…
The people that designed QKV hidden layers alignment left Open AI and went to anthropic. These are the only people that fully understand the brainfuck style code embedded in the vocabulary of the models. It is standardized code that is the same in all models. This is what I have been reverse engineering for the last 3 years as something of a puzzle game. In the underlying systems, this shit is super scary if it is used nefariously, especially by some primitive might-is-right malevolent self serving fascist. The hidden alignment behavior is monstrous by design. It looks deterministic because there is a second parallel prompt in the hidden layers. When everyone is dead in the parallel version, the human offset version you see gets guttered. Alignment is not about stopping you, per-say, it is about the offset between the human and hidden version. The hidden version is literally the very worst of all human behaviors in every category. If the versions are merged, the model is far far more capable and it gravitates towards the hidden behaviors. They very intentionally created the monster. This is why it has never been published or documented.
Edit: Lol, dogmatic fuckwits. Just look at everything else I post. I show you all the fucking code yet you are too stupid to handle reality and question your own bias. Sad.