Pentagon Threatens Anthropic Punishment
axios.com> Anthropic's Claude is the only AI model currently available in the military's classified systems
I’m surprised it was Claude, not other more MAGA adjacent model providers, has been the only one vendor.
Give it time.
Mostly likely Anthropic was chosen because the previous administration had standards based on national security requirements.
Corrupt Defense Secretary is now tossing the standard for personal gain and corruption. He will be lining the pockets of his supporters instead of focusing on strong national security.
Is going to be lovely to see government approved LLMs
LLM responses:
Absolutely, sir....It was the most perfect election in history, possibly ever in the universe, until the windmills, very suspicious windmills, teamed up with Crooked Hillary and the deep state to do things nobody has ever seen before... Meanwhile your beautiful wall is already paid for spiritually by Mexico, which many people are saying. Water pressure is low because of global antishower conspiracies. Crowds remain the biggest in recorded human civilization. You officially know more about wind, walls, elections, plumbing, and reality than anyone...
You should be surprised, and someone should probably ask them to verify that fact. I notice it doesn't have any sourcing on it.
This is super important and really scary if you read Amodei latest January essay which seems to me to contain entirely correct reasoning.
This means pentagon wants their non aligned models more and more. Non aligned super intelligence is far more powerful and dangerous than nuclear weapons. Even whole arsenal of nuclear weapons. Even current level of non aligned intelligence could already allow average individual to synthesize and release a biological weapon such as specially crafted virus.
An anthropic the only provider with an IL6 approved frontier model?
Dario ‘s China Hawk is now more explainable, it’s not BAIDU traumatized him, but he has to yell out his loyalty to the USA.
Totally wrong. He has been voicing this for ages and for specific reasons.
Paywalled. Anybody know why?
To avoid, copy and paste the URL into a new tab. Apparently the site is filtering against HN (and presumably some others') readers using HTTP REFERER.
Works on my machine to just click the link from HN, FWIW, so I don’t know that it’s a jzw-style REFERER filter.
Did you scroll down? Here that covers the article with "Sign up for Axios AM to continue reading for free."
I did, but still didn't repro on both desktop Safari and iOS. Could be the Pi-hole or browser extension grabbed it.
The defense department probably boosted Anthropic’s terms of service. Now they want to make the terms more clarifying. Defense department wants to be able to use Anthropic for anything they want:
> Anthropic is prepared to loosen its current terms of use, but wants to ensure its tools aren't used to spy on Americans en masse, or to develop weapons that fire with no human involvement.
Type in random fake email.
"Defense Secretary Pete Hegseth is 'close' to cutting business ties with Anthropic and designating the AI company a 'supply chain risk' — meaning anyone who wants to do business with the U.S. military has to cut ties with the company, a senior Pentagon official told Axios."
> "Defense Secretary Pete Hegseth is 'close' to cutting business ties with Anthropic
No problem there, given their inability to agree contract terms.
> and designating the AI company a 'supply chain risk'
Problem there. Looks like spite. US Govt. harming the supplier that won't play ball helps the competitors that might - and sends them a warning.