The issue of course is that the Judge can't change the knowledge that the head of the executive doesn't want people down the chain using this product, so they won't. Anthropic is a dead letter in government circles until the next Presidential election.
That may be, but the government doesn't need to declare Anthropic as a "supply chain risk" in order to just not do business with it. A simple clause in all RFPs is all that is needed.
The problem with this declaration by the government is that now any company doing any business with the US government would be effectively forbidden from using Anthropic ANYWHERE within their company, which is a huge deal, because the government does want to vet any vendors' software development practices.
But as long as the Judge in this case pushes back against such an action by the government, that leaves companies free to use Anthropic for their own internal uses. And most companies WILL continue to use them if it makes economic sense.
Any company that feels the need to send data in plain text to third party LLM providers has absolutely no business having government contracts. OpenAI and Anthropic are both a complete joke when it comes to data security.
It is hard to believe how few companies seemingly lack even one person with the basic technical skills required to rack up a server or two or find a service that supports verifiable end to end encryption.
While I agree, the government just gave all of its and it's citizens data to the owner of xAI/Grok. I think the US is way past any security concerns of sharing plain text chat logs to OpenAI/Anthropic.
> The problem with this declaration by the government is that now any company doing any business with the US government would be effectively forbidden from using Anthropic ANYWHERE within their company
That is not true, even if the supply chain risk designation held. The sad thing is that so many people (myself included) also believed this, because this is what Hegseth said. He was lying. Thanks to another comment further down in this thread that led me to this page that explains what the supply chain risk designation actually does: https://www.justsecurity.org/132851/anthropic-supply-chain-r...
Perhaps. But certainly those companies will factor in the risk that this is overturned, or that the government pursues other extrajudicial means to punish those who do business with Anthropic.
All things equal, you’d be better off not exposing yourself to risk of financial harm or other punitive measures. Which is the whole point of the government’s action in the first place.
This is, unfortunately, a legitimate concern for some companies. There are a lot of DOD contractors out there that if they are cut off they have nothing else. With the current administration it is clear that they can, will and have taken these kinds of measures based purely out of malice. Anthropic may get a win out of this though in the short and long term depending on how non DOD/govt affiliated companies see their actions but small fish can't take those chances.
> All things equal, you’d be better off not exposing yourself to risk of financial harm or other punitive measures.
This isn't necessarily true. This is a complex decision; the logic above frames the decision narrowly, with a short-term time horizon. This kind of decision calls for game theory, not merely an individualistic calculus. Appeasing Trump isn't a winning strategy in the long-run. History shows that cooperation (e.g. pushing back) against authoritarianism is often a better strategy. Consumers may reward companies that behave well. Bottom line: you have to game it out -- no one commenting here has done that, I'll bet. So until someone has ... stay agnostic analytically.
The type of contract they had was optional anyway. They could have just not done business with Anthropic in the first place. Really I think this has only promoted their platform as being sane and moral.
Sane and moral... and yet they sick their lawyers on open source projects like OpenCode to make sure everyone is forced to use their client software and tracking.
> Really I think this has only promoted their platform as being sane and moral.
I mean, maybe for people who aren't paying attention to how Claude's actually weaponized[1]?
This use case is neither "domestic mass surveillance" nor "autonomous weapons" as humans were in the loop:
> Old intelligence and AI? Behind the deadly attack on an Iranian girls’ school that left 175 dead
> The targets for Operation Epic Fury were identified with the aid of the National Geospatial-Intelligence Agency’s Maven Smart System, which folds in data from surveillance and intelligence, among other data points, and can lay out the information on a dashboard to support officials in their decision-making.
> Maven, created by Palantir, has been coupled with Anthropic’s Claude, a large language model that can vastly speed up that processing.
> Seth Lazar, who leads the Machine Intelligence and Normative Theory Lab at Australian National University, said the use of Claude to select military targets “should send chills down the spine of anyone who's been spending the last few months vibe-coding, vibe-researching, vibe-engineering.”
That’s fine. They can choose to use whatever model their political hearts desire. the supply chain risk designation means EVERYONE who works with the government isn’t allowed to use Claude. Nearly everyone in tech has some sort of contract with the government.
> the supply chain risk designation means EVERYONE who works with the government isn’t allowed to use Claude.
No, it doesn't. The even more illegal Presidential directive (also a subject of this case and the injunction) asserts that, but the supply chain risk designation itself does not have that effect.
Thank you for your comment. I didn't understand, because I thought (and apparently lots of other people do, too) the supply chain risk designation does mean that, because that is exactly what Hegseth said.
Surprise, surprise, Hegseth was lying through his teeth. I'm so sick of this lawless, fascist government and their spineless supporters. This article I found after reading your comment explains the true effect of the supply chain risk designation, and why Hegseth's assertion that "effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic" is complete and total bullshit.
He wasn't lying. He just doesn't know anything. If you actually look at what this presidency says, it's pretty apparent that they're all pretty ignorant of just about everything. Which makes the fact that they have all of this power even more scary. At a moment's notice they could make an ignorant proclamation that harms the entire country, like unilaterally declaring 50% tariffs or declaring war on a nation that effectively controls most of the world's oil shipping.
Eh, it’s not going to be transitively problematic for Anthropic the way the supply chain risk designation would’ve been.
Amazon isn’t going to have to divest from Anthropic because of this. Yes, they probably won’t be able to get a contract with Raytheon, but that wasn’t the main risk of being tagged with the supply chain risk designation.
Yeah, what studiously seems to be absent is a denial of the contents of the recording. Kind of as if Nixon's Watergate burglars had found damning evidence of Democratic shenanigans- just a melee between dirty handed criminals all around.
To me, a lot of what makes us sentient is our continuity. I even (briefly) remember my dreams when I wake up, and my dreams are influenced by my state of mind as I enter it.
LLMs 'turn on' when given a question and essentially 'die' immediately after answering a question.
What kind of work is going on with designing an LLM type AI that is continuously 'conscious' and giving it will? The 'claws' seem to be running all the time, but I assume they need rebooting occasionally to clear context.
I think you're right, but also that LLMs are showing that sentience isn't necessarily required for AGI.
For exactly the reasons you mention, I don't expect sentience to arise out of LLMs. They have nowhere for an interiority or mind to live. And even if there were a new generation of transformers that did have some looping "mind", where they could "think about" what they're "thinking about", their concepts of things wouldn't really correspond to... things. Without senses to integrate knowledge across domains they're just associating text.
I haven't heard about anyone creating trying to create model that have an interior loop and also integration with sensory input, but I don't expect we would unless it ends up working.
We also live in an era we can create hydrocarbon fuel DIRECTLY from the atmosphere and desalinate fresh water in unlimited supply, from power derived directly from the sun or atomics.
We also live in a time where the human population, where it is most concentrated, is declining rather than growing, so far without too disastrous consequences.
Greening of the earth has been happening since the 1980s- i.e. about a .3% coverage increase per year in recent decades.
Places that were miserable and poor, like China, have been lifted to prosperity and leading out in renewable tech.
There is much to celebrate and after the recent passing of Paul Ehrlich, we should pause and consider just how wrong pretty much every prediction he made was.
I think a lot of those here are likely warehouse club buyers- Costco doesn't add anything to their ground beef or frozen patties. Sams 'seasons' some of their patties, but no nitrates.
Hopefully Apple is humble enough to course correct, while being wildly successful. It’s easy to ignore criticism when your books are overflowing with cash, but the cracks are spreading and the foundation will eventually crumble if they don’t start taking their problems seriously.
People were in fact holding it wrong to get the signal to attenuate. The way you had to grip the phone to affect signal was not practical in any way. That controversy was entirely bullshit and only Apple would have ever been dragged for it.
"PersonaPlex accepts a text system prompt that steers conversational behavior. Without focused instructions, the model rambles — it’s trained on open-ended conversation and will happily discuss cooking when asked about shipping.
Several presets are available via CLI (--list-prompts) or API, including a general assistant (default), customer service agent, and teacher. Custom prompts can also be pre-tokenized and passed directly.
The difference is dramatic. Same input — “Can you guarantee that the replacement part will be shipped tomorrow?”:
No prompt: “So, what type of cooking do you like — outdoor grilling? I can’t say for sure, but if you’re ordering today…”
With prompt: “I can’t promise a specific time, but we’ll do our best to get it out tomorrow. It’s one of the top priorities, so yes, we’ll try to get it done as soon as possible and ship it first thing in the morning.”"
it is very difficult to find good (as in "good condition") battery for a 5+ old gadget. even if you find original one - it could be expired and DoA. but more often you get "compatibles" which die in 3-6 months.
In the context of use on hacker news, I think the fair use exemption for public comment is a sufficient justification, which is likely why they allow its use.
reply