|
| ▲ | sashank_1509 2 days ago | parent | next [-] |
| Yes, anthropic wanted that power through a legal agreement. Not by spying on the pentagon , or training their AI model to lie to them etc, which seems more appropriate for supply chain risk. The government in this case can just cancel its legal agreement with Anthropic and move on, which was always its expected move. Trying to unilaterally destroy Anthropics business for a contractual disagreement is not fair and I’m glad the judicial is pushing back. |
|
| ▲ | Dylan16807 2 days ago | parent | prev | next [-] |
| Anthropic wanted power over what the government would do with the servers Anthropic runs. That's not weird and that's not being a risk. It's normal business negotiation. |
| |
| ▲ | charcircuit 2 days ago | parent [-] | | This is not about the usage restrictions in the agreement. As shown by OpenAI's agreement getting similar restrictions. It is about Anthropic's attitude in how Anthropic has the ultimate power in its usage. If RTX started demanding that they should be the ones to decide who Tomahawk missiles can be used on when they are launched. And RTX said that the government should file a support ticket to appeal a decision, then I would not be surprised that such actions could lead to considering them to be a supply risk. Even if it was just part of "business negotiation." It is the mindset that the other company has which clearly is showing signs of risk. | | |
| ▲ | SpicyLemonZest 2 days ago | parent [-] | | What the court found is that Anthropic demanded no such thing. The government lied and claimed that they did as part of their attempt to punish Anthropic. | | |
| ▲ | charcircuit 2 days ago | parent [-] | | It was a preliminary injunction. The purpose of such an injunction is not to establish what actually happened. We will need to wait for this to progress further to learn more information about what actually happened. | | |
|
|
|
|
| ▲ | joe5150 2 days ago | parent | prev [-] |
| The military can work with someone else's product or use a bit of their trillion-dollar budget to come up with their own. |
| |
| ▲ | charcircuit 2 days ago | parent [-] | | Yes, they can work with someone else. Marking them as a supply risk is one way they can avoid using them and instead use someone else for their needs. So now it seems like we are in a limbo where the government knows that Anthropic is a risk to work with, but they can't official put them on a list that states that. | | |
| ▲ | joe5150 2 days ago | parent | next [-] | | They actually don't need to sanction Anthropic as a "supply risk" in order to not contract with them, and doing so is obviously an insane overreach. | |
| ▲ | MeetingsBrowser 2 days ago | parent | prev [-] | | Can you explain how Anthropic is a risk to work with? | | |
| ▲ | charcircuit 2 days ago | parent [-] | | Put simply, the military should have to ask Anthropic for permission each time it needs intelligence. Time is of the essence for the military and having to argue over these things in the moment is not good. These things should be figured out ahead of time and or properly reviewed afterwards. Working with such a company that demands to embed themself into this process with the power to deny any request is too much power. The risk to the military for companies working with Anthropic is that they can get delays or outages when there shouldn't be which can jeopardize time sensitive operations. |
|
|
|
