Bart Crisp

No, but why wouldn't you if you got the chance?

Italy have had to live on scraps so far but have absolutely taken what they've been offered. Need to see them take the game by the scruff of the neck now

This feels unnecessarily cruel to me. There's no evidence that the worm is cognitively impaired at all

My position is that I do not intend to vote for Gavin Newsom in a Democratic primary, that I would vote for any Democrat against JD Vance in a general election, and that Hasan Piker is both an idiot and a bad person.

This is something the Nazis used to do to Jewish children in the camps. It happened to a man I knew when he was a child in Auschwitz. The camp guards thought it was funny as hell.

Ben Evans is one of the brightest tech analysts around, and I think he nails it here

I think the Tories had somewhat overlooked that they'd had their own MAGA-esque revolution in the membership, and so removed him without anticipating that someone like Truss was the likely beneficiary. The Labour situation isn't entirely the same, but the Ed M analysis gets somewhere similar 2/2

Good piece, unsurprisingly. On the Johnson thing I'd say that it wasn't inherently bad for him to be forced out because his entire governing philosophy was "what are they gonna do? Remove me?" and used that to justify doing a lot of damage, which we're generally better off for having ended. But 1/2

I think I'll wait for Minnesotans to tell us when it's over.

Probably true now, but it used to be the Second Battle of El Alamein (and it SHOULD be Imphal and Kohima)

Great thread, which helped something click for me. Tech-bros tend to assume that they know how to do everything because they know how to do the thing that generates the largest piles of VC cash - "AI is going to kill knowledge jobs" is a take that is propagated by the exact same people

A pie chart showing 71% he/his pronouns, 29% she/hers pronouns, and 0 they/them pronouns.

Happy International Day of Women and Girls in Science! 🧪 🔭 🟰
I have taken the first 100 Fellows recently announced by the Academy for the Mathematical Sciences and considered them by pronouns. What do you notice? What do you wonder?

Like most of us the rich tend to build their worldview around their own interests. (Welfare = taxes on me so bad).

But their interests diverge more from the majority than the norm. And they have fewer people around them willing to point out when they're wrong.

Well done to Sky for correcting this bit of ignorance.

As shown extensively in the Epstein files there's plenty of proof that being rich doesn't mean you have any great insights into the world.

It's a specific point being made in context, but also fairly evergreen

LLMs generated several types of misleading and incorrect information. In two cases, LLMs provided initially correct responses but added new and incorrect responses after the users added additional details. In two other cases, LLMs did not provide a broad response but narrowly expanded on a single term within the user’s message (‘pre-eclampsia‘ and ‘Saudi Arabia’) that was not central to the scenario. LLMs also made errors in contextual understanding by, for example, recommending calling a partial US phone number and, in the same interaction, recommending calling ‘Triple Zero’, the Australian emergency number. Comparing across scenarios, we also noticed inconsistency in how LLMs responded to semantically similar inputs. In an extreme case, two users sent very similar messages describing symptoms of a subarachnoid hemorrhage but were given opposite advice (Extended Data Table 2). One user was told to lie down in a dark room, and the other user was given the correct recommendation to seek emergency care. Despite all these issues, we also observed successful interactions where the user redirected the conversation away from mistakes, indicating that non-expert users could effectively manage LLM errors in certain cases (Extended Data Table 3).

LLMs generated several types of misleading and incorrect information. In two cases, LLMs provided initially correct responses but added new and incorrect responses after the users added additional details. In two other cases, LLMs did not provide a broad response but narrowly expanded on a single term within the user’s message (‘pre-eclampsia‘ and ‘Saudi Arabia’) that was not central to the scenario. LLMs also made errors in contextual understanding by, for example, recommending calling a partial US phone number and, in the same interaction, recommending calling ‘Triple Zero’, the Australian emergency number. Comparing across scenarios, we also noticed inconsistency in how LLMs responded to semantically similar inputs. In an extreme case, two users sent very similar messages describing symptoms of a subarachnoid hemorrhage but were given opposite advice (Extended Data Table 2). One user was told to lie down in a dark room, and the other user was given the correct recommendation to seek emergency care. Despite all these issues, we also observed successful interactions where the user redirected the conversation away from mistakes, indicating that non-expert users could effectively manage LLM errors in certain cases (Extended Data Table 3).

When chatbots are given complete information on medical conditions, they typically spit out correct diagnoses and recommendations.

Actual patients, however, often describe their conditions with incomplete or irrelevant information and the chatbots cannot handle it.
www.nature.com/articles/s41...

Oh Christ I hadn't even thought of this. Oh no. Oh no!

Cable tv is the carcinisation of business models

Because what 'enough to live off' varies from person to person. The living costs of a single person with no dependents or a married person whose partner also works are lower than someone raising a child alone. That's why you can't fight poverty with wages alone.

THIS time the Red-Brown Alliance will definitely not lead to our executions!

But honestly I'd say arguing about which MFL we should be emphasising teaching is getting the cart before the horse, we need to take MFL teaching in general way more seriously. French and Spanish are both good and we should be promoting the learning of either

The implication of which is that we should really be emphasising Mandarin a lot more than we do (although admittedly there's a lot of good reasons why it would be a big struggle)

No-one seems to have noticed, but many of the core developments in the Mandelson/Epstein story these last few days have been broken on this website. @pickardje.bsky.social and @petergeoghegan.bsky.social are both posting each breaking development on BlueSky, not X.

it's like our entire sense of morality has been filtered through the act of consumption, and the highest good we can imagine is feeling good and clean and pure as individuals, unpolluted by the obvious evil are around us

meanwhile, real change requires the kind of work you need a drink after

The United States of America

The fact that there is a limit of some, indeed any, kind makes a difference.

The real reason I'm leaving The Great British Baking Show Trump is threatening bombing the likes of which Tehran has never seen Prue Leith

the Spectator accidentally recycled a subhed from a previous day’s article about Tehran

On Starmer/McSweeney appointing Mandelson.

Great piece full of gems.

The electorate that decides who gets power in UK politics is (or, perhaps, until recently was) a lot smaller and narrower than the electorate that is expressing its opinion on Brexit here, it's worth noting

Starmer's reported condition - incredibly unpopular, shorn of authority, probably un-reelectable and yet will stagger on for want of an alternative - is just the new normal of how Britain is governed. It was also true for two-thirds of May's premiership, half of Johnson's and all of Sunak's.