Das Pentagon hat gerade eine erschreckende Nachricht an KI-Unternehmen gesendet

The line is too blurry to implement. If you are against your product being used for their stated limits you just cant work with the DoW. Its impossible to isolate distribution and usage within programs so its really either in or out. Even the managerial work done in DoW is all in support of weapons development, surveillance, or violence enabling.

Sadly, we know what happens when you take a knife to a gunfight. We know Russia/China/North Korea will rush head long into AI weapon craft, not matter how we limit dense/fast AI chips. As noble as it sounds – It is not noble, it is foolish. We have seen what happened when iron met bronze, or when tanks met horses(WW1) and later on when Panzers met WW1 tanks in WW2, and now we see how drones are defeating tanks.
Hiding your head in a hole will not protect you, as it did not(as rumored) protect the Ostritch.
So the Pentagon is right = no knives to gunfights for me!!

The government not buying a technology because the manufacturer puts their own limits on it is totally fine. Ultimately, in a well-functioning government, the political process dictates those limits not private companies.

As always, the real problems here are:

(1) This administration can never be trusted with anything

(2) The use of totally unrelated government authority to punish companies (not buying from them is fine, but threatening unrelated regulatory action would be very very bad)

(3) AI is being used in deeply reckless ways

Why can’t they do what every other defense contractor does and take the money and then take forever yo not deliver anything functional….saying the program
Needs more as more money?

yo leave Claude alone!!

“A spokesperson for Anthropic denied that such a conversation had taken place in a statement to Axios, saying the company had “not discussed the use of Claude for specific operations with the Department of War.”

What about conversations with the **Department of Defense?** You know, the actual name of the department.

„Our nation requires that our partners be willing to help our warfighters win in any fight“ – this is referring to them using autonomous weapons without human intervention. I cannot imagine a situation where it’s a positive thing for non-human weapons to work by itself unless they planned on using those weapons in a way that humans would find distasteful or disagree with. It sounds like the DOJ is trying to plan an attack against US Citizens or bypass a certain level of restrictions on the person to person illegal order rejection for nukes potentially.

This is absolutely terrifying.

Enable fascists to get more power and they own you. Congratulations silicone valley, you supported this

I hope people are starting to realize the dangers of unethical AI use.

If I were Anthropic or Palantir, I would sever ties. Take the hit, and start applying their CV knowledge to something that benefits people.

All AI companies should have Anthropic’s stance. This should not be terrifying. Stand on morals, assholes.

“You’re absolutely right, those were innocent American citizens. I’ll enter Plan mode and come up with a revised set of instructions for the bomb-civilians skill that will ensure this doesn’t happen again.”

“Mr Hegseth, you are mistaken.
We are Anthropic. I suggest you go down the street to our competitor, Misanthropic.”

It’s going to happen and we are fxxx

Dying agency scratching and clawing to stay relevant

Shut all this shit TF down. Honestly.

We’re fucking ourselves over SO HARD

Access to the military will be a double edged sword for tech companies. Once you’re integrated into secretive systems they’ll take whatever they want under the guise of national security.