“It sounds insane and happened in Italy in the 80s or 90s” are astonishingly good criteria for this. False flags, fascists, communists, assassinations, organised crime, massive corruption. There’s even a subplot featuring terrorist Freemasons en.wikipedia.org/wiki/Propaga...
product image from amazon where one of the answers is "to be fully transparent as an AI"
Help I found this journal on amazon and the prompts in the image were clearly filled out by AI
We trained our business AI on thousands of failed companies doesn’t sound like a great selling point.
Geoff Mulgan April 14th, 2026 The myth of STEM only growth holds back the UK
‘Why did the USSR, with its fantastic scientists and engineers and heavy investment in STEM, nevertheless stagnate?’
Excoriating from @geoffmulgan.bsky.social on the undervaluing of social science and humanities research
blogs.lse.ac.uk/impactofsoci...
How do we make invention and innovation central to the UK's future?
In the latest Options for the UK essay, Nesta’s Joe Owen proposes a 'British Nobel Prize' - a future-focused challenge award to reignite our culture of invention and solve major societal problems of our time.
How could it work? ⤵️
Would anyone be able to tell?
Their follow-up post nails it: “semantically ungrounded but (largely) syntactically valid and hyper-convincing text”
boxobarks.leaflet.pub/3mj42airv3s2o
(To be clear I’m not disagreeing with the post, it makes the point too, I’m just saying that for me the generally low semantic content of AI text is its defining feature, more than its distinctive turns of phrase)
This is kind of interesting on some of the “tells” of AI writing.
For me what’s distinctive about AI text is more general though - a sense of reading something where the words are unmoored from underlying meaning. I try to grab at it and … there’s not enough there to catch hold of.
Writing a law is not something for which there is a technological solution. It is not a perfectible process, it is a moral act that requires belief and responsibility. It is a process of debate. As MPs, advisers and lobbyists know, the real business of our constitution happens in the background – in emails, notes, agendas. If everyone involved is asking the same software to condense emails and write replies, if they are reading research and updates and memos composed by the same software, that software increasingly assumes the power of the people who previously did the thinking. Reading is thinking, and writing is thinking, and thinking is power. And when the inefficiencies of human thought, deliberation and opinion are cleared aside, we are left asking: who is in charge?
This is a very good article. www.newstatesman.com/technology/2...
As the Artemis astronauts return from the Moon, it feels relevant to repost this… on Guyana’s small but important contribution to the Apollo 11 landing ojusher.medium.com/the-stories-...
“vibe McKinsey-style reports” is an extremely funny sequence of words
No Brit should be allowed to post anything naming a US Supreme Court justice unless they can name all of the members of the British Supreme Court.
And so in that respect, AI slop isn’t all that different from old style handcrafted artisanal small-batch slop?
For sure. But maybe a lot of human communication has been slop all along?
Anyway this reminds me a) that semiotics is a field that exists that I ought to do some light reading about, and b) I ought to re-read Foucault’s Pendulum.
All these things vaguely wave a hand in the direction of meaning, through context and aesthetics, rather than really having semantic content. In think there’s something really fascinating (not in a very good way!) about how humans communicate.
It’s been clear for a long time that many people think images are just a vibe (hence stock photos), numbers are a vibe (BS impact stats and KPIs) and diagrams a vibe (see most PowerPoints). So it shouldn’t have come as a big surprise when LLM use revealed that, to many, written info is a vibe too.
The Hollywood space movie aesthetic is the 70s NASA aesthetic though, not the other way round. The Apollo footage is unreal, so much stuff filmed from multiple angles in 70mm film for posterity etc. There’s a reason this film looks so good, and it’s not cgi. www.imdb.com/title/tt8760...
This reminds me of when Sam Altman announced that OpenAI’s new model was “great at creative writing”, and posted an example that was exactly what someone who had never read a novel in his life would think creative writing was like
Interesting, I did not know that. Is that about the fibres themselves? Or the tech for encoding information in light? (If the latter - isn’t that downstream of semiconductor improvements?)
Literally the only exponential technology ever is semiconductors, and even that is hitting hard limits now.
The irony is that going slower - more deliberately designed - is going faster. Embedding AI in a human-centered (rather than state or corporate centered) way will make it more productive, equitable and less likely to generate political backlash.
We never learn.
giftarticle.ft.com/giftarticle/...
Comic Sans. Papyrus. Clashing colours. Love it.
Bonus. This is what I get when I tell it to make the slide "really really really really ugly", and of course it looks much cooler.
11th generation
Weird filigree background. Horrible cursive font.
10th generation
Jaunty asymmetric effect. Three crowns instead of one.
9th generation
The AI sparkle has been replaced with a crown. The old font has been retired and replaced with a bolder one that can take thicker bevels, carry more gold and cast deeper shadows.
8th generation
The lighting has got even moodier. The letters are now not just in 3D, but have a bevelled effect.
7th generation
The background now looks slightly uneven, as if it were covered in real gold leaf.
6th generation