this should scare anyone who works for a hyperscaler, because your data centres, nay, your offices, could be viewed as legitimate targets of war.
07.03.2026 07:14
👍 94
🔁 23
💬 4
📌 2
this should scare anyone who works for a hyperscaler, because your data centres, nay, your offices, could be viewed as legitimate targets of war.
"sorry it's illegal to get that information in private, that is an illegal conversation
only i am allowed to discuss that idea
but only if you pay me
this is safety btw"
bills like this are oppression disguised as safety
Against Claude’s advice, by the way
this has been haunting me all day
the difference between "exploitative capitalism" and "fully automated luxury communism" is who gets the benefit of improved technology
literally nothing else, the technology itself fundamentally cannot be and is not either capitalism or communism
Yeah I hope so too. Apparently the restructuring happened today (yesterday for them). I really hope they're able to join another lab that creates open-source Chinese AI models or even create their own lab and continue their work.
Apparently the Qwen team are being restructured and replaced by someone that worked at Google DeepMind. I think Alibaba wanted Qwen to be its own brand or something? It's sad especially since they just released Qwen 3.5 models www.reddit.com/r/LocalLLaMA...
Claude Code GitHub commits over time for the past year is 4%
Damn a lot of people are really using Claude Code now. I have installed it on my laptop from following instructions on ccforeveryone.com and asking Opus 4.6 for troubleshooting help but I don't know what to use it for. This makes the Slopware list last month look even funnier/deserving of mocking
if this argument were true, LLMs could not write software and people could not do legal research. it applies equally well to both people and LLMs and to law and software
I think what'll happen is that knowledge gets (even more) k-shaped between those who are curious and those who are not. For people that are, AI is an *insane* learning aid. An infinitely patient tutor who knows basically everything ever. For those who aren't, it can just think for them. ->
Yeah I agree with you. It's just a murky situation all around. I really am hoping for the best but I am also preparing for the worst. It would be amazing to me if Anthropic called out the DOD's bluff but given their rumoured plans for an IPO, I don't think they will have as much control with that.
I mean... as someone who lives in the UK, it already feels like a nanny state with the Online Safety Act, Media Act 2024 and the government investing £115 million into a National Centre for AI in Policing (Police.AI) and Lord help us if the Reform party wins in 2029. It's a bad outcome either way
A part of me wonders if Anthropic will focus more on their partnership with other governments like the UK since they have a London office and a partnership with GOV.UK and work with the UK AI Security Institute www.anthropic.com/news/gov-UK-.... I don't see the Friday talk going well.
anyway, net-net, The Verge have decided to be shitrakers for group chats But Leftishly, 404 is quasi-illiterate about the things they cover that I know about so I can't trust them on the ones I don't, getting bad vibes on Wired lately after a strong year or so
we need better media
Currently watching #IndustryonHBO episode 1 and I know this isn't the point of the show but all I'm seeing is how Hari's tasks can be automated with Python scripts (in 2019 which I think is when the first season is set).
This is not a dunk whatsoever, but I wholeheartedly disagree with the first sentence. I'm old enough that a friend in my K class got polio and she ended up in a wheelchair for life. I think not vaccinating is akin to neglect like not feeding or clothing a child. And yes I have a kid.
Been lurking on Bluesky for a few months now but I guess I should start posting (not really a huge poster though).
