A lot of people on Bluesky sneering at this. But Russia has no interest in helping Iran, apart from the fact they're a key regional ally, they sell them Shahed drones, and that a prolonged war and with it rising energy costs help the Russian economy.
Mangled missile fragments purporting to be from the deadly strikes that hit a naval base and elementary school in southern Iran on Feb. 28 bear the markings of an American cruise missile, according to an analysis by The New York Times.
Because Die Welt is not enough (HT Nick Mathiason)
That would presumably be this www.theguardian.com/politics/201...
If you're using your phone or laptop and thinking everything feels super glitchy and janky, you're not alone. But it's not necessarily AI's fault. My latest for @inc.com pinpoints what's going on www.inc.com/chris-stokel...
?
Intriguing final point in this that tbh I'd like to see explored more.
4. US government designates Anthropic as a supply chain risk Extremely low (<1%). Anthropic is a US-founded, US-headquartered company with strong ties to the US national security community and a reputation for safety-focused AI. There's no plausible near-term pathway to this.
2. US–Israel joint airstrikes on Iran to assassinate head of state Low (~3-5%). Both countries view Iran as a major adversary, and Israel has conducted targeted assassinations of Iranian figures. But a joint operation specifically targeting the head of state would be an extraordinary escalation — essentially an act of war — with enormous geopolitical consequences. Covert Israeli operations are more plausible than overt joint strikes.
one of the fun things about AI is using it as a "just woke up from a coma" simulator
Every time you say you haven't been somewhere in the UK I am more than mildly surprised. Just assume you've been everywhere in search of content.
I do not believe a former tabloid employee did not intend that pun...
I have seen this film, and I do not remember this scene...
I think if anything it will remind a few people he didn't immediately get on board, which will probably be slightly good for him?
You cannot talk like this about any other minority in America.
And if it was any other president’s adviser who said something like this about any other minority, it’d be the instant end of their presidency.
Very, very normal place for a byline. Also, casting aspersions on the actual editor without any evidence too!
What Al Jazeera doc are you referring to?
Oh, lol, didn't need to ask my previous question!
What is the specific corruption mechanism you are referring to here? Who are you alleging is paying or otherwise reimbursing her for these (IMO dumb and morally bad) policies and statements?
This is pretty grotesque from Badenoch. But also deeply politically weird: the Iran strikes are already unpopular and will become even more so. Why lock herself into a stance voters will hate?
As a bonus: the initial strikes are clearly illegal. Joining in defence to counterstrikes isn't.
Source describes the failed Pentagon-Anthropic talks: through the end, the Pentagon wanted to use Anthropic's AI to analyze bulk data collected about Americans (Ross Andersen/The Atlantic)
Main Link | Techmeme Permalink
OpenAI posted the terms of the deal. Reveals that it absolutely does allow for domestic surveillance. EO 12333 is how the NSA hides its domestic surveillance by capturing communications by tapping into lines *outside the US* even if it contains info from/on US persons.
openai.com/index/our-ag...
Screenshot of New York Times article: “OpenAI Reaches A.I. Agreement With Defense Dept. After Anthropic Clash”
Screenshot of text from article: “Under the deal, OpenAI agreed to let the Pentagon use its A.I. systems for any lawful purpose, a term required by the Pentagon. But OpenAI also said it had found a way to ensure that its technologies would adhere to its safety principles by installing specific technical guardrails on its systems.”
My major takeaway from the last year of reporting on generative AI chatbots is that safety guardrails can fail when conversations run long and that everyone who works in this space knows that
www.nytimes.com/2026/02/27/t...
Good for deleting and then posting separately to own up to the mistake tho! Good practice in action.
Yes, global electricity consumption by data centres may double by 2030. I wish the FT had added some more proportionality from the same IEA report: “However, in the wider context, a 3% share in 2030 means the data centre share in global electricity demand remains limited.”
www.ft.com/content/47da...
(I don't disagree with the overall point, but I do think that we should "give someone a cookie" when they do the right thing, cos the whole point of that is to train them to do the right thing more often....)
I am no Anthropic fan boy, but isn't this an example of it being quite hard not to do the morally dubious thing, with quite major consequences, and they've chosen not to do the morally dubious thing? Kinda feel like not giving credit simply encourages the rest to go full morally dubious...
Wrote about how the incredibly thin-skinned leadership of Palantir sued a small Swiss publication for publishing a deeply researched article that was only mildly embarrassing for the company (about why the Swiss gov't keeps ghosting them).
This actually exists I'm afraid.
