Pasta

[Pasta @DashBoost_Pasta]

@sama The contract doesn't say "will not be used" full stop; it says "will not be used in any case where law or policy [prohibits]". My understand on Anthropics concerns were that there is no law that prohibits the gov't from automated killing, and policy can change. What don't I get?

[Alan Rozenshtein @ARozenshtein]

These are NOT meaningful redlines. For example it only prohibits autonomous weapons “ in any case where law, regulation, or Department policy requires human control.” But the relevant safeguard against autonomous weapons is a DOD directive that Hegseth can change at will!

Also the surveillance redline is about “unconstrained” surveillance of “private” information. But what about “slightly constrained” surveillance of private information, or unconstrained surveillance of “public” information? Those are both potentially very dangerous forms of mass surveillance!

[Aakash Gupta @aakashgupta]

OpenAI says their Pentagon deal “has more guardrails than any previous agreement for classified AI deployments, including Anthropic’s.”

Let’s examine that claim.

Anthropic tried to write explicit contractual prohibitions binding the Pentagon regardless of future legal changes. The Pentagon refused. Anthropic told ABC News the compromise language was “paired with legalese that would allow those safeguards to be disregarded at will.” That’s what blew up the deal.

OpenAI’s contract “explicitly references the surveillance and autonomous weapons laws and policies as they exist today.” Axios confirmed the restrictions “reflect existing U.S. law” and “the intention was not to invent new legal standards.” This tells you everything about the structural difference. One company tried to create new binding obligations. The other pointed at obligations that already exist and put them in a contract.

A reference clause and a prohibition clause look similar on a press release. In a courtroom they’re completely different instruments.

OpenAI says if the government violates the terms, “we could terminate the contract.” Anthropic had that same right. The government just demonstrated what happens when you exercise it: supply chain risk designation, federal agency ban, every defense contractor barred from doing business with you. The termination clause is real. The ability to invoke it without getting Anthropic’d is not.

OpenAI says cloud-only deployment prevents autonomous weapons because “this would require edge deployment.” That assumes the military will never build a low-latency connection between a cloud API and an edge weapons system. Any defense contractor who has built a kill chain knows that boundary is an engineering problem with a timeline measured in months, not a physics constraint.

The forward-deployed engineers with security clearances are the strongest part of OpenAI’s case. Anthropic deployed through Palantir. OpenAI is putting its own people inside classified environments. That’s a real structural upgrade. But “in the loop” is doing enormous work. Real-time query approval is a guardrail. Monthly log audits are a rearview mirror. OpenAI hasn’t specified which one.

One contract tried to say “you cannot do this.” The other says “we’ll watch to make sure you don’t do this, and also the law already says you can’t.” A contract lawyer would never confuse those two things. And after this week, neither should anyone else.

[Daniel Litt @littmath]

Not a lawyer, but I don’t understand how this, from OAI’s contract with the DoW, enforces their apparent redline “No use of OpenAI technology to direct autonomous weapons systems.”

[NotedallaSfera @TalkingMusicz]

As always… you just twist the narrative to suit your agenda. Professional GASLIGHTERS. You absolutely kept the “all lawful use” , and subordinated the usage of AI to applicable laws. Anthropic rejected this specifically because the democratic law doesn’t yet appropriately regulate AI, which means you simply gave the government the free pass to operate in the grey area we are today. You signed exactly what Anthropic rejected. You have no credibility whatsoever, but this had already been apparent for months.

[Barack Obama @BarackObama]

The killing of Alex Pretti is a heartbreaking tragedy. It should also be a wake-up call to every American, regardless of party, that many of our core values as a nation are increasingly under assault.

[Bill Clinton @BillClinton]

Over the course of a lifetime, we face only a few moments where the decisions we make and the actions we take will shape our history for years to come.  This is one of them.

[Pasta @DashBoost_Pasta]

Got a response over email:

It's absolutely ridiculous to block an account and provide no justification.

[Pasta @DashBoost_Pasta]

My @coinbase @CoinbaseSupport account has been restricted for weeks now, and the support has been atrocious. Absolutely terrible.

This is why people are moving away from exchanges like coinbase and kraken, and moving towards DEX's

A thread. 🧵

[Pasta @DashBoost_Pasta]

Legacy brokerages have their problems, but anytime I have an issue with fidelity or vanguard, I call a number and the problem is solved.

@coinbase, do better. Your customers deserve transparency and real support.

End of thread. 🚨

[Pasta @DashBoost_Pasta]

@gthomaskerr @coinbase @CoinbaseSupport I don't, and no-one should, hold on exchange. Just valuable to have an onramp from cash to crypto.