I have become more and more leery of "leaked" internal letters or memos from companies and the government given how they have tended to play out in the public. However, I do not think that is the case when it come to this letter given how pointed and how it really harms any sort of reproachment between the U.S. Admin and Anthropic.

I am trying to find full access to the letter but have not been able to yet. Numerous people have confirmed parts of the letter though and Anthropic is going full scorched earth. In the letter CEO Dario Amodei made the following comments:

• "The real reasons DoW and the Trump admin do not like us is that we haven’t donated to Trump (while OpenAI/Greg have donated a lot),"
• "We haven’t given dictator-style praise to Trump (while Sam has)."
• Amodei claimed the company had risked the fallout of the Admin because it "supported AI regulation which is against their agenda, we’ve told the truth about a number of AI policy issues (like job displacement), and we’ve actually held our red lines with integrity rather than colluding with them to produce “safety theater” for the benefit of employees (which, I absolutely swear to you, is what literally everyone at DoW, Palantir, our political consultants, etc, assumed was the problem we were trying to solve)."
• He called how OpenAI has messaged the clash as "mendacious" and "an example of who they really are, and I want to make sure [everyone] sees it for what it is."
• He went after OpenAI's safeguards calling them mostly safety theatre and "mostly do not work" adding "false that OpenAI’s terms meaningfully protect them against domestic mass surveillance and fully autonomous weapons."
• Finally he wraps it up with "Near the end of the negotiation the DoW offered to accept our current terms if we deleted a specific phrase about 'analysis of bulk acquired data,' which was the single line in the contract that exactly matched this scenario we were most worried about. A lot of OpenAI and DoW messaging just straight up lies about these issues or tries to confuse them."

Pick really any of the points and you can easily see how this would be problematic for the Admin to reverse course anytime soon or Anthropic to change its mind. The last point though is one that I think shows a critical flaw in what Anthropic wanted to be able to do.

DoW offered to accept our current terms if we deleted a specific phrase about 'analysis of bulk acquired data,'

What this means at least on the government side is something that is just never going to happen. Anthropic wanted to be able to look through the data that was collected, given to it, etc. by the DoW. Remember this was on classified networks as well and essentially Anthropic was asking to open up classified networks and the data in them to more people and that would be irresponsible for the U.S. Government to do. As soon as something critical to safety leaked people would be see self inflicted vulnerabilities like this and be appalled.

When the House was looking to officially partner and allow for Members and Staff to use ChatGPT, Gemini, Copilot, etc. My friend who was the lead in developing the firewalls to protect House data and sensitive data was locked in a battle for over a year with these companies over the House's data. Initially all of them wanted access to everything, the inputs the outputs, just everything and that alone was a huge no go. It was only later on that OpenAI agreed to let a firewall be put around their products to allow us to be able to go to ChatGPT and others ended up following.

One company did not come to an agreement and that is the company we are talking about right now. Anthropic wanted access to all the data and wanted to be able to use it to train. Obviously that was not something the House could agree to and so Claude is not available to the House. I have been bringing in my own personal computer for over a year just to be able to use it for researching topic and searching for legislation.

Anthropic's CEO might have bit off more than he can chew with this one and shareholders could easily come pounding down his door. Given the companies seemly uncompromising stance of access to all data that is used on its products the enterprise value, which Anthropic has been an early mover on, is going to end up crushed. There is no reason for a company to trust Anthropic with proprietary information. The quickest way I see this getting resolved is with Dario Amodei being forced out. I agree with him in a perfect world what Anthropic wants would be possible but this is not a perfect world and when you want to be a go to top tier company you have to realize this and not live in a fantasy world.