Interesting news from work: it turns out only one local authority in England runs monthly Client Level Data (CLD) extracts...

...which also runs queries compiling Adult Social Care Outcomes Framework (ASCOF) measures from them, based loosely on NHS GitHub scripts...

...and DHSC are interested...

/whistles innocently...

(Albeit I'll have to rewrite the scripts to look at a standard extract rather than the custom one we use which converts text formatted date fields into proper date fields)

IKEA España are apparently now distributing empathy instruction manuals with Blåhaj!

https://youtu.be/OwgvshnbgM8

The prime minister has resigned.

See also: 2024, 2022, 2021, 2019, 2018, 2016.

Noting what's happened to a certain water feature in DC...

Make Algae Great Again? 😈

"How does the spring not run out almost immediately?" "We pull it back REALLY far."
https://xkcd.com/3244/

Birmingham: taking No Overall Control to the extreme...

With 101 seats in total, no two parties combined have enough seats to get over the 51 vote threshold to pass motions, so ensuring not just cross-party but cross-ideology support is needed.

Oh, and we still have central government appointed Commissioners until October 2028, scrutinising all major spending decisions and potentially blocking those they claim are unaffordable (such as a proposed resolution to the refuse strikes involving £16k payouts to every affected worker).

Another Reform defector returns to the Conservatives, claiming "the vast majority of candidates are indeed racist, misogynistic, and bigoted", while they have "also have been significantly frustrated and dismayed by the failure of the Reform Party’s leadership to tackle this issue in any meaningful way, and their attempts to instead try to brush it under the carpet or cry foul play." 🤦

Politics is indeed beyond parody... 🤷

https://www.standard.co.uk/news/politics/london-reform-candidate-defects-tories-georgia-david-racism-b1168085.html?utm_source=firefox-newtab-en-gb

Amid the chaos of the news stories of dubious veracity today, a former legal professional has released his annual double-entendre-laden video...

https://youtu.be/vY8azAWTiYM

Donald's got another made-up award to add to his collection: the "America First" award, courtesy of the Republican National Congress...

https://www.independent.co.uk/news/world/americas/us-politics/trump-republican-fundraiser-speech-award-b2945851.html

Reform's candidate vetting is as strict as ever...

...sorry, I meant to say as hopeless as ever...

https://www.theguardian.com/politics/2026/mar/20/reform-uk-suspends-scottish-candidate-stuart-niven

Brum News: Unite the Union have been fined £265,000 by the High Court for breaching the depot blockade injunction (+£170,000 towards the council's legal costs, with a possible further £90,000 under consideration).

https://www.birmingham.gov.uk/news/article/1752/unite_fined_265000_for_breaching_high_court_injunction_at_council_waste_depots

They do like being descriptive, with field names of 30-50 characters being common... (!)

The final output tables are also written as a giant string (so the table names can be passed as variables) then EXEC'd, to add complexity to the 700+ line script...

Currently trying to coerce an NHS SQL script to calculate waiting times between requests and assessments / services into something that will work on our system.

Good news: the script now runs without error.
Bad news: the output tables are empty.

So now to exclude the last table from the script, and do a count(*) on every temp table created en-route, to work out which filter's screwed up, then why...

Nigel's latest gambit (when he's not flying half way across the world to avoid being anywhere near Clacton...): investing in a Bitcoin reserve company run by Kwasi Kwarteng (of Liz Truss mini-budget fame)...

https://www.standard.co.uk/business/business-news/kwasi-kwarteng-nigel-farage-bitcoin-london-reform-uk-b1274025.html

Spotted on FB:

The Iranian religious leadership could seriously confuse the Americans by electing Donald Trump as honorary supreme leader (with a deputy providing day to day great leadership).

Jackson's research is great for exploring this, because we get to see abstract synthetic images that very strongly stimulate the AI to see... whatever it "wants" to see.

Often, the results are recognizable. The image with oddly shaped pink blobs does sorta resemble flamingos. But there are also many examples where the AI fixates on some small detail of color or texture, and becomes convinced it's seeing something totally implausible.

This is relates to "adversarial examples", another great way to see this.

With real images, it seems like the AI "sees" like we do. But as soon as we venture beyond its training data, the illusion is broken, and it feels a bit like a parlor trick. Clearly AI doesn't see like we do.

This is a great practice for AI generally: seek out the edge cases where the model fails. This breaks the spell of "general intelligence" and gives us a clearer idea of what's actually happening inside the black box.

(3/3)
#science #ai #generativeart

LLM image generators can make a picture of anything you ask for. The results often look pretty good at first glance. They're generic and the details are usually off, but folks overlook that easily.

This hides something important: these models can't see like we do. The main limitation is how they're trained. We show them millions of pictures, paired with text descriptions.

The problem is, humans don't describe images literally. We might say "a picture of a dog playing frisbee" but we didn't mention the setting, the composition, or the squirrel in the background.

Most of what's there visually is unsaid. The model sees those pixels, but they're just "stuff that goes along" with the text. Dogs play in parks, so the AI learns that dogs have green backgrounds.

This is why it's so hard to control an image generator. It isn't intentionally placing all those objects and choosing their attributes, it's just extra fluff that seems to "go with" what you asked for.

(2/3)
#science #ai #generativeart

My lab mate, Jackson Dean, has been doing some really fun research into image generation.

Unlike the common AI-generated images that mash together stolen artwork to make something sorta photo realistic, he's producing abstract art that's entirely novel. The general idea (inspired by innovation engines) is to generate an image from scratch, then ask a vision / language model what it sees. He generates lots of images with different descriptions, and refines those images to more closely resemble their descriptions.

Not only is he making some really cool generative art, but he's learning something about what "novelty" is and how to produce it in a computer.

Beyond that, though, I'm fascinated because it gives a window into the strange way computers "see" images.

(1/3)
#science #ai #generativeart

NEWS! Donald Trump's biggest critic now his own 2013 Twitter account https://newsthump.com/2026/03/02/donald-trumps-biggest-critic-now-his-own-2013-twitter-account-3/

Good morning.