Hear me out:
12.03.2026 14:16
👍 65
🔁 12
💬 1
📌 0
Hear me out:
I just can't parse "advanced hindsight". It makes my head hurt.
An article “More than one million foreign nationals are receiving benefits” (March 17, 2025) referred to the number of “foreign nationals” receiving benefits, and gave the rate of benefit claims for various nationalities. In fact, the particular DWP data on which the article was based recorded the nationality of claimants when registering for a National Insurance number, rather than the nationality at the time they were claiming benefits, so that some claimants might have since acquired UK citizenship, and would not be foreign nationals at the time they were claiming benefits. This correction has been published following an upheld ruling by the Independent Press Standards Organisation. Advertisement
Fully *one year later*, the Telegraph has been forced (by @ipso.co.uk) to admit that it was lying about foreign nationals claiming benefits.
"Extractive education is a bad idea"
Exhibit #29183747372
Actually, this on a banknote
Our new report with @cnn.com found that 8 in 10 AI chatbots typically assisted with planning violent attacks, including school shootings.
While safeguards exist, many companies aren’t using them.
Lawmakers must require safety-by-design before these tools fuel more real-world harm.
More ⤵️
I'm suing Grammarly over its paid AI feature that presented editing suggestions as if they came from me - and many other writers and journalists - without consent.
State law requires consent before someone's name can be used for commercial purposes.
www.wired.com/story/gramma...
"we're just making it up and you'll pay us for it"
there was also a Stanford study that found that a surge in AI slop by lazy people actually created significantly more work for others who had to decipher and correct it
this "I'm being efficient" (but not really) vibes well with the American obsession with artifice and appearing productive
True, I guess I'm just seeing the enormous amounts of cash flowing into these AI companies and thinking how many amazing studies that money could fund instead.
Who did this?
(Well done)
Ok agree, but I also feel there should be some quid pro quo back for profit-driven use of public data. I see it as akin to the large scale copyright theft that's already happened re training LLMs.
But h/t to @pengzell.bsky.social, out I am probably talking nonsense bc cmd line Claude is local!
Cool. I'll shut up then
IDK! If so then great! Wonder if that means it's calling python and R locally then?
Public service announcement: Flo will be on BBC Radio 1 tonight between 8 and 10, with Henry
open.spotify.com/track/4wgzZy...
Same same though? The CMD line version is for code nerds?
It's being held on US servers for the work to be done by Claude is what the app said when asked.
As to public research: sure, absolutely, but that doesn't mean giving away very valuable data - paid for by the public - to companies that use it for profit.
Well, the machine says it doesn't retain the data but it does go to US servers
Data goes to the US.
Ok but no no no. Genuinely think that UKDA and funding councils need to act on this because otherwise tens of millions in research ££ will be given away for free
Well, if eg Anthropic does end up with loads of very-expensive-to-collect UK data they can pay a LOT of money for it 😉
This is very fun
Does that then mean Anthropic have BES data and it can be used by others??
"If you are tracking the impact of AI on labor markets, CEO statements are the worst way to measure it." Amen, @marthagimbel.bsky.social! @hamiltonproject.org #AIPlusWork
Labour MP Charlotte Nichols reveals she was raped as an MP as she opposes the governments jury trial reforms
Every Labour MP should listen to her incredibly powerful and brave statement in parliament today
If we hadn't sold north sea to private companies we'd have a lot MORE options
Chutzpah. That's what you mean.
"Cambridge. Oh yeah. There's that OTHER university there too... University of Cambridge"
