• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Trump Orders Government to Stop Using Anthropic After Pentagon Standoff

P

philb2

2[H]4U
Joined
May 26, 2021
Messages
3,369
Writing on Truth Social, Mr. Trump used harsh words for Anthropic, describing it as a “radical Left AI company run by people who have no idea what the real World is all about.”

Shortly after Mr. Trump’s announcement, and 13 minutes after a Pentagon deadline, Defense Secretary Pete Hegseth designated the company a “supply-chain risk to national security.” The label means that no contractor or supplier that works with the military can do business with Anthropic. Later on Friday, OpenAI, the maker of ChatGPT, said that it had reached an agreement with the Defense Department to provide its A.I. technology for classified systems.

The Anthropic designation was all but unheard-of, legal experts said. It stripped an American company of its government work by using a process previously deployed only with foreign companies the United States considered security risks.


https://www.nytimes.com/2026/02/27/...5926&user_id=c383821527c441214d07ce6e4a6ba12a
 
I would think it's vital that a single person be able to activate any Ai equipped military equipment, quickly, alone, and without over site or legal interference.

I mean not to democracies, but authoritarian regimes would really need it or they wouldn't be able to be authoritarian.
 
techie81 said:
Government contracts aren't a right and I am sure people wouldn't bat an eye if it was the other way around.
Click to expand...
Yes, but this threat is about just not giving them a contract, but about blacklisting them and saying nobody in your supply chain can work with them. Granted, I think it is a hollow threat but the goal is simple ... threaten to destroy them and hope they cave. Saying that they are a risk to national security is a bit of a bigger deal than just not giving them a contract.
 
I wonder how that will affect their Claude model. I tried many many models. I consider Claude to be much better that ChatGPT and Gemini. It is sharp/focused and just works. It can generate and edit/modify office files in seconds following your instructions to the letter. OpenAI ChatGPT and Google Gemini don't even come close when it comes to editing files automatically for you. I feel bad for Anthropic.


https://www.techrepublic.com/articl..._86b8j27yn&utm_term=daily-tech-insider-active
 
tunatime said:
It's always been that way but it's less worse then it was under Biden. You just didn't hear about the strong arm because the other side was doing it so it was ok.
Click to expand...
No you didn't hear about it because Biden didn't send our tweets in all caps letting the world know of his absolute authority
 
sram said:
I wonder how that will affect their Claude model. I tried many many models. I consider Claude to be much better that ChatGPT and Gemini. It is sharp/focused and just works. It can generate and edit/modify office files in seconds following your instructions to the letter. OpenAI ChatGPT and Google Gemini don't even come close when it comes to editing files automatically for you. I feel bad for Anthropic.


https://www.techrepublic.com/articl..._86b8j27yn&utm_term=daily-tech-insider-active
Click to expand...
Politics aside, Claude beats the snot out of Gemini and ChatGPT for coding/scripting/misc office tasks I’ve thrown at it. I like playing around with local LLMs on my server gear, but for the few practical use cases (AKA those worth spending money on…) I’ve also found Claude to be better, sometimes a lot better. Strongly dislike Anthropic’s convoluted usage tiers vs pricing though.
 
techie81 said:
Government contracts aren't a right and I am sure people wouldn't bat an eye if it was the other way around.
Click to expand...
You’re correct, but the administration is doing more than that, they are declaring them a supply chain threat which means any other company in the US that works with the federal government cannot use Claude. That’s insane. The supply chain threat thing is typically used for foreign companies with clear foreign intelligence ties. The administration is blatantly abusing the office to be punitive to a US based company because that company isn’t fucking insane and doesn’t want their AI to be used to basically recreate the movie Wargames or Terminator 2.

People should be absolutely terrified that we are now at a scenario where idiots in government want to completely automate our defense with AI. I’m far more worried about the outcome of that than the threat of China now.
 
sram said:
I wonder how that will affect their Claude model. I tried many many models. I consider Claude to be much better that ChatGPT and Gemini. It is sharp/focused and just works. It can generate and edit/modify office files in seconds following your instructions to the letter. OpenAI ChatGPT and Google Gemini don't even come close when it comes to editing files automatically for you. I feel bad for Anthropic.


https://www.techrepublic.com/articl..._86b8j27yn&utm_term=daily-tech-insider-active
Click to expand...
No good deed goes unpunished.
 
SamuelL421 said:
Politics aside, Claude beats the snot out of Gemini and ChatGPT for coding/scripting/misc office tasks I’ve thrown at it. I like playing around with local LLMs on my server gear, but for the few practical use cases (AKA those worth spending money on…) I’ve also found Claude to be better, sometimes a lot better. Strongly dislike Anthropic’s convoluted usage tiers vs pricing though.
Click to expand...
"coding/scripting/misc office tasks I’ve thrown at it."


Exactly, it did save me hours of work. ChatGPT and Gemini will either take ages to do it or deliver screwed up work that is completely useless, and no matter how you try to fix it with superb prompt engineering, it still won't work.
 
ERipley said:
ahhhh yes, the good ol' Trump good/Biden bad excuses... God bless Merica
Click to expand...

I wouldn't give a shit if they said "If you don't follow our rules you can't have the contract". That's perfectly normal and pretty much par for the course with any contract. If you don't want to play ball the way they want you to then why should they pay you?

The problem crops up when they threaten to designate someone a national security risk when they're obviously not even close to being one. If they tried they'd be dragged into court and have to prove it which they'd obviously fail at. That means they're probably only making stupid bombastic threats like they always do and will quietly drop the whole thing once people are distracted by the next big bombastic announcement or threats. Iran is doing a good job of that right now so they can move on from threatening Anthropic to doing something else dumb.
 
The big question is how the conditions that OpenAI puts on are different from Anthropic's. They look very similar from here.
 
uOpt said:
The big question is how the conditions that OpenAI puts on are different from Anthropic's. They look very similar from here.
Click to expand...
Yeah. The only reason Sam did that was the big uptick in people leaving the service. I think that number jumped like 300%? Some not insignificant shift and he panicked and had to back down and admit he didn't think it through.
 
And the groveling begins:

https://www.anthropic.com/news/where-stand-department-war

1772770654680.png
 
As we wrote on Friday, we do not believe this action is legally sound, and we see no choice but to challenge it in court.
Click to expand...

Good, take em to court, likely wont go anywhere, but defamation seems appropriate to go after when the president of a country labels your company as such which could have financial impacts to those who believe Trumps words...
 
MrGuvernment said:
Good, take em to court, likely wont go anywhere, but defamation seems appropriate to go after when the president of a country labels your company as such which could have financial impacts to those who believe Trumps words...
Click to expand...

The big financial hit to Anthropic will be investors who are nervous about putting money into them if the government tries to ban them. If Anthropic can prove that Trump's tantrum cost them many billions in investor dollars they might make out like gangbusters on the taxpayer dime.
 
I am sure this decision had nothing to do with the fact that Anthropic was leaking a crapload of data to Chinese national companies…..
Companies who utilized exploits to gain full developer access…
 
Lakados said:
I am sure this decision had nothing to do with the fact that Anthropic was leaking a crapload of data to Chinese national companies…..
Companies who utilized exploits to gain full developer access…
Click to expand...
I think it was just a case of Anthropic being obstinate and unwilling to "bend the knee".

My understanding of the Anthropic/China thing from last month was that there was an organized effort (from Chinese AI firms) to do prompt injection to parse out how models were trained. More like disclosing hints of the recipe to train another Claude model rather than leaking any private data. Outlets picking it up as "DATA LEAKED!!" was just the typical hyperbolic, shitty reporting by tech-illiterate mainstream news... It's more boring to say "companies attempting industrial espionage for trade secrets".
 
SamuelL421 said:
Politics aside, Claude beats the snot out of Gemini and ChatGPT for coding/scripting/misc office tasks I’ve thrown at it. I like playing around with local LLMs on my server gear, but for the few practical use cases (AKA those worth spending money on…) I’ve also found Claude to be better, sometimes a lot better. Strongly dislike Anthropic’s convoluted usage tiers vs pricing though.
Click to expand...

Ehh. I tried Claude, actually just because the Government went after them and I wanted to throw them a few bucks.

A big flaw I found was Claude is it's like 50 first dates. Every context is walled from every other context. The implications are that you have to teach it everything every time you open a new context window. In addition for the pro version, the context window is embarrassingly small, I filled it in 10 hours.Yeah you can spend $100 to get a bigger context window, but it has to be greater than it is to get that. I can't imagine this model is good for war.

So for simple tasks it may be better if you don't mind teaching it over and over. For complex tasks, I use gemini and had used chatGPT but the latter failed me on many things.
 
Last edited:
schmide said:
Ehh. I tried Claude, actually just because the Government went after them and I wanted to throw them a few bucks.

A big flaw I found was Claude is it's like 50 first dates. Every context is walled from every other context. The implications are that you have to teach it everything every time you open a new context window. In addition for the pro version, this context window is embarrassingly small, I filled it in 10 hours.Yeah you can spend $100 to get a bigger context window, but it has to be greater than it is to get that. I can't imagine this model is good for war.

So for simple tasks it may be better if you don't mind teaching it over and over. For complex tasks, I use gemini and had used chatGPT but the latter failed me on many things.
Click to expand...
Yeah my experience had been similar, Anthropic's usage tiers complete BS. There's basically 2 plans: Pro (half-assed teaser for the full functionality) and Max (legitimately better at scripting/code than anything else I've tried). I've heard Gemini is solid too but I haven't tried it on a paid plan yet. Supposedly Google's "Ultra" plan gives access to all their skunkworks -type LLM tools - I'd love to play around with that but it's something crazy like $250 a month.
I can't imagine this model is good for war.
Click to expand...
I can't see any of these commercial models having any real utility (out of the box) for DoD purposes. I'm sure they would have to build/train a model from scratch for a given military purpose.
 

“Claude’s consumer growth surge continues after Pentagon deal debacle​


Claude’s daily active users are on the rise on mobile devices, as are its new app installs, following the company’s fallout with the Pentagon. After Anthropic CEO Dario Amodei refused to allow the government to use its AI systems for mass surveillance of Americans or to power fully autonomous weapons, the AI model provider behind Claude was marked as a supply-chain risk.
However, Anthropic’s stance led many consumers to favor the model, data suggests.”

https://techcrunch.com/2026/03/06/c...-surge-continues-after-pentagon-deal-debacle/
 
What if this was all a big brain move to get the very people Trump wants to get data on to migrate to their platform?

Sign a back-room deal, have the Anthropic people say "we denounce Trump and everything he wants to use our platform for!", the anti-Trump people flock to the platform, start collecting their data...

In a few months, a 3'rd party company nobody cares about with some boring ass name like "AI Research LTD" comes in and starts doing everything the Trump administration was going to do with the platform, and then they just sell that work back to the government...
 
Lakados said:
Sign a back-room deal, have the Anthropic people say "we denounce Trump and everything he wants to use our platform for!", the anti-Trump people flock to the platform, start collecting their data...
Click to expand...
While devious I think you are over estimating the amounts of people using AI outside of what's pre-built into their phone.
 
sfsuphysics said:
While devious I think you are over estimating the amounts of people using AI outside of what's pre-built into their phone.
Click to expand...

I think it's also underestimating the Government's ability/capability to collect data on everyone/anyone outside of AI/Claude/etc - lest people forget Prism/Snowden/etc
 
sfsuphysics said:
While devious I think you are over estimating the amounts of people using AI outside of what's pre-built into their phone.
Click to expand...
I think you overestimate most government officials and their ability to put together a coherent thought...
Democrat think tanks, interns, and various staff members, speech writers, legal analysts, writers for TV talk shows...

Maybe it's just competing doorbell or security surveillance platforms, etc.
 
SamuelL421 said:
Yeah my experience had been similar, Anthropic's usage tiers complete BS. There's basically 2 plans: Pro (half-assed teaser for the full functionality) and Max (legitimately better at scripting/code than anything else I've tried). I've heard Gemini is solid too but I haven't tried it on a paid plan yet. Supposedly Google's "Ultra" plan gives access to all their skunkworks -type LLM tools - I'd love to play around with that but it's something crazy like $250 a month.

I can't see any of these commercial models having any real utility (out of the box) for DoD purposes. I'm sure they would have to build/train a model from scratch for a given military purpose.
Click to expand...
I think there are special versions/models for governments or defense, not what we normal users have access to. Both with OpenAI and Anthropic.
 
The regime literally said it was because they wanted safeguards disabled as did Anthropic.

So now it's because most Glorious, Omnipotent, Generous, and Infinitely Compassionate leader is playing 45D chess and it's all a ruse.

At least Jim gave his meat-bots their merch for free, and the Kool-Aid.
 
You must log in or register to reply here.
Back
Top