I do think the “all of humanity” stuff is a little overblown, but this is legitimately dumb and dangerous and will get a ton of innocents killed + allow the military to dodge accountability. Letting ChatGPT potentially run an autonomous weapon with zero oversight is phenomenally stupid and the tech is nowhere near reliable enough to be able to pull off the kind of precision and decision making military campaigns require, which is what Marcus is saying

this truly is the dumbest timeline huh

Incredibly ballsy move to keep using their tech after you literally branded them a supply chain threat and implied you would take legal action against them, but that’s this administration for ya

(they did say there would be a six-month phase out period after which if Anthropic still didn’t comply, they’d force them to, but still)

Late on Friday, rival OpenAI, which is backed by Microsoft, Amazon and others, announced its own deal to deploy technology in the Defense Department’s classified network. CEO Sam Altman on X said the Pentagon shared its principles for human responsibility over weapon systems and for having no mass U.S. surveillance.

So safe to assume that’s complete bullshit right? Altman literally got fired from his company for being a massive liar, and as I already mentioned OpenAI already got caught working with the US Government to create a surveillance system. His announcement on shitter already got community noted as well:

"Government officials have contradicted Sam’s claim, saying OpenAI will allow the DoW to use their models for “all lawful purposes”. This could allow for uses Anthropic refused to engage in, namely mass surveillance tools and weapons systems with no human oversight.”

oh yeah SamA’s statement was definitely PR-adjacent (OpenAI already got caught working with the US government and the people behind Discord’s age verification to create mass surveillance) but Trump’s threats against Anthropic are definitely real

(edit: https://youtu.be/zZ98DPIp0a4 source for the OpenAI surveillance thing)

yeah I suspected Scam Altman was full of shit (he already got caught working with Trump and the people behind Discord’s age verification for mass surveillance), but Trump’s threats against Anthropic are genuine, “major criminal consequences” imply that Anthropic workers may be in real danger of getting arrested (edited to add link)

UPDATE 2: The plot has thickened, the other AI companies are getting involved now

The US vs Anthropic plot has thickened considerably

so I guess that’s that for now.

I mean I wouldn’t be surprised if the DoD actually does follow through with their threats against Anthropic because that’s perfectly in character for this administration

UPDATE: Anthropic’s responded and is standing their ground on no mass surveillance and autonomous weapons

deleted by creator

Just read this piece by Gary Marcus on this topic. Jesus fucking Christ this administration sucks so fucking bad. As a matter of fact ‘sucks bad’ is an understatement for the sheer amount of damage this administration is causing

Latest batch of Anthropic nonsense dropped http://youtube.com/post/UgkxzQmoMujNPQ6rLLGfrXxI67pGVSJrpu9J

If I had to list every single worldwide problem right now, Trump would be connected to at least 80% of that list

the US Government slides further into technofascism via throwing chatbots into the military https://garymarcus.substack.com/p/code-red-for-humanity you don’t hate this administration enough

but right now it doesn’t feel like LW is a place where I can collaboratively make intellectual progress on this very important topic.

Maybe that’s because people don’t feel comfortable when you say you should teach an AI race/iq pseudoscience???

Literally everything I’ve heard about OpenClaw is it fucking up and being a security risk. A sensible person would realise this means it’s unreliable and shouldn’t be left to take care of important tasks (or really just tasks in general), but AI boosters are typically not sensible people. And the guy who made it got a job at OpenAI, so this is far from the end…

this is like the fourth time an AI agent has completely deleted something important (I remember an article about an AI deleting all of a scientists’s research) How many more times does it have to happen before people stop using AI to look after something important???

sharing this channel’s posts are the equivalent of shooting fish in a barrel but http://youtube.com/post/UgkxoSpDpLNEr9WawVXnl5Mlw4NeQ6-XsLjl this really just feels like an excuse to repost that METR graph. also wtf is the graph on top

it’s always the Elon Musk fans isnt it.

and on the topic of Futurism articles on Elon Musk: https://futurism.com/future-society/court-trouble-jury-hates-elon-musk

one word: LMFAOOOO