Something doesn’t make sense here. His tweet claims he has exactly the same restrictions that Anthropic had.

skissane an hour ago | parent | next [-]

This tweet (from Under Secretary of State Jeremy Lewin) explains it:

https://x.com/UnderSecretaryF/status/2027594072811098230

https://xcancel.com/UnderSecretaryF/status/20275940728110982...

The OpenAI-DoW contract says "all lawful uses", and then reiterates the existing statutory limits on DoW operations. So it basically spells out in more detail what "all lawful uses" actually means under existing law. Of course, I expect it leaves interpreting that law up to the government, and Congress may change that law in the future.

Anthropic wanted to go beyond that. They wanted contractual limitations on those use cases that are stronger than the existing statutory limitations.

OpenAI has essentially agreed to a political fudge in which the Pentagon gets "all lawful uses" along with some ineffective language which sounds like what Anthropic wanted but is actually weaker. Anthropic wasn't willing to accept the fudge.

	▲	qdotme 44 minutes ago \| parent \| next [-]
		Well, or just the possibility of future-proofing the agreement in favor of the US government, as well as walking back the slippery slope of „no autonomic lethality” and „no mass surveillance”. The former, grants the Congress the ability to change the definition of all „lawful use” as democratically mandated (if the war is officially declared, if the martial law is officially declared). The latter, is subtle. There can exist a human responsibility for lethal actions taken by fully autonomous AI - the individual who deploys it, for instance, can be made responsible for the consequences even if each individual „pulling of a trigger” has no human in the loop (Dario’s PoV unacceptable). Similarly, and less subtly, acceptance of foreign mass surveillance, domestic surveillance (as long as its lawful and not meeting the unlawful mass surveillance limits!) seems to be more in the Pentagon’s favor. Whether we like it or not, we’re heading into some very unstable time. Anthropic wanted to anchor its performance to stable (maybe stale) social norms, Pentagon wanted to rely on AI provider even as we change those norms.
	▲	PakG1 an hour ago \| parent \| prev [-]
		Because the US government has such a great track record on ensuring that this kind of stuff is only done legally with the utmost integrity. /s

▲

Jensson 2 hours ago | parent | prev | next [-]

Sam probably told them they can renegotiate those restrictions in a year or so when the drama has died down.

▲

patcon an hour ago | parent | next [-]

yeah, something shady. i don't trust sam at all.

i once ran into someone in london in 2023 who was doing their thesis on AI regulation. they had essentially ended up doing a case-study on sam. their honest non-academic conclusion (which they shared quietly) was that they were absolutely terrified of sam altman.

fear is one of those signals we ought to listen to more often

▲

m3kw9 an hour ago | parent | prev [-]

Is not shady, the systems are not ready for that kind of task esp autonomous hunting. Is smart negotiations, plus Sam would have used the Anthropic situation against them saying you can’t designate all AI top American AI companies supply chain risk etc. it’s complete idiocy the would do that anyways

	▲	qdotme 36 minutes ago \| parent [-]
		Ready at what level, though. The subtleties are what matters. It’s well established that belligerents can use mines, to separate the tactical decision of deploying for purposes of area denial; from the snap-second lethal decision (if one can stretch that definition) to detonate in response to an triggering event. Dario’s model prohibits using AI to decide between enemy combatant and an innocent civilian (even if the AI is bad at it, it is better than just detonating anyways); Sam’s model inherits the notion that the „responsible human” is one that decided to mine that bridge; and AI can make the kill decision. How is that fundamentally different in the future war where an officer might make a decision to send a bunch of drones up; but the drones themselves take on the lethal choice of enemy/ally/no-combatant engagement without any human in the loop? ELI5 why we can’t view these as smarter mines?

▲

labrador an hour ago | parent | prev | next [-]

This is a actaully a government bailout of OpenAI. Investors gave it a bunch of money earlier knowing this was going to happen. Greg Brockman is a major Republican donor for 2026. Nice for OpenAI.

▲

ddtaylor 2 hours ago | parent | prev | next [-]

PR spin/lying while behind closed doors agreeing to it. What's hard to understand about OpenAI lying?

Altman publicly claimed he had no financial stake in OpenAI to emphasize his mission-driven focus. In 2024 it was revealed that Altman personally owned the OpenAI Startup Fund.

In May 2024, actress Scarlett Johansson accused Altman of intentionally mimicking her voice for ChatGPT's "Sky" persona after she had explicitly declined to work with them.

When OpenAI’s aggressive non-disparagement agreements were leaked, which threatened to strip departing employees of all their vested equity (potentially millions of dollars) if they criticized the company, Altman claimed he was unaware of the "provision."

▲

gritspants 2 hours ago | parent | prev | next [-]

My theory is that they both went through normal procurement processes. At some point, one of Palantir's forward deployed sales agents slapped someone's arm at the golph course and said, yes we can automously kill with our AI agents. Anthropic, having little to do with the kind of 'AI' in a use case that made sense for, declined.

	▲	jaco6 2 hours ago \| parent [-]
		[dead]

▲

straydusk 2 hours ago | parent | prev | next [-]

I know the reaction to this, if you're a rational observer, is "OpenAI have cut corners or made concessions that Anthropic did not, that's the only thing that makes sense."

However, if you live in the US and pay a passing attention to our idiotic politics, you know this is right out of the Trump playbook. It goes like this:

* Make a negotiation personal

* Emotionally lash out and kill the negotiation

* Complete a worse or similar deal, with a worse or similar party

* Celebrate your worse deal as a better deal

Importantly, you must waste enormous time and resources to secure nothing of substance.

That's why I actually believe that OpenAI will meet the same bar Anthropic did, at least for now. Will they continue to, in the same way Anthropic would have? Seems unlikely, but we'll see.

▲

moralestapia 2 hours ago | parent | prev | next [-]

Makes 100% sense.

They said yes to the same thing.

	▲	karmasimida 2 hours ago \| parent [-]
		Dario is being ruled out due to ideological standing Makes perfect sense

▲

anigbrowl an hour ago | parent | prev | next [-]

You really think someone would do that, just go on the internet and tell lies?

https://knowyourmeme.com/memes/just-go-on-the-internet-and-t...

▲

Tadpole9181 2 hours ago | parent | prev | next [-]

Well tweets aren't legally binding, so chances are he's just outright lying so they can have their cake (DoD contracts) and eat it too (no bad PR)

	▲	jkaplowitz 36 minutes ago \| parent \| next [-]
		> Well tweets aren't legally binding There's nothing in general about a tweet that makes it any more or less legally binding than any other public communication, and they certainly can be used in legally binding ways. But sure, a simple assertion to the public from the CEO of a privately held company about what a separate contract says is not legally binding - whether through tweet, blog, press release, news interview, or any other method.
	▲	sudo_cowsay an hour ago \| parent \| prev \| next [-]
		companies like saying things that makes it look like they aren't doing anything bad but then they decide to do exactly what they said they wouldnt e.g. google project maven, microsoft hololens (military), and much much more
	▲	nurettin 4 minutes ago \| parent \| prev [-]
		This is so funny to me. Especially since Elon musk had to buy Twitter due to his tweets.

▲

foobarqux 2 hours ago | parent | prev [-]

No, the difference is that the government agrees to no "unlawful" use as determined by the government.

Anthropic said that mass surveillance was per se prohibited even if the government self-certified that it was lawful.