AI for leaders and policymakers | Author of the AI Sovereignty Tracker machinepowerindex.org | Bylines in @spectator, @unherd and more | elliotleavy.com
London
- Tweets 1,357
- Following 1,120
- Followers 133
- Likes 299
Pinned Tweet
If you missed my remarks on the case for Sovereign AI, here’s my talk from Coordinating British Progress on Friday.
Thanks again both @aeronlaffere and @lilygeidelberg for organising and to @WalkerMarcus for generously hosting us in the extraordinary setting of St Bartholomew the Great.
machinepower.substack.com/p/…
1
1
10
1,599
I put together a pub-first manifesto a year back and every AI tool suggested this provision. Makes you think 🤔.
🍻New: Pubs that are tied to large companies like Greene King or Admiral will have new rights to serve a guest beer of their choice
The new 'Guest Beer' provision will be put into the England & Wales Pubs Code
28
Elliot Leavy retweeted
BREAKING: China orders government agencies to develop plans to combat mass job displacement from AI.
162
680
6,489
478,422
Starting on my 2027 trends report, so I went back to last year’s to see what landed. Nine of the twelve calls have held so far and the other three were partly right, which I’ll take. Thanks to everyone who read it first time round.
Read the full redux (receipts included), below.
On next year's report, plan is to eat some of Evan's lunch. Stay tuned.
1
1
66
"The major advances in civilization are processes which all but wreck the societies in which they occur." A. N. Whitehead
25
Every country that isn't the US or China now has two ways to stay in the AI race: leverage or theft. Expect a lot more of the second. My piece for @unherd 👇🏻.
‘Frontier AI now carries the same strategic weight as the atomic bomb once did.’
@ElliotLeavy on the OpenAI espionage case 👇 buff.ly/GvkYpxE
1
1
66
Feels like a stopgap to me.
If we think about McLuhan, a new medium doesn't just add itself to the old one — it reshapes the entire cognitive environment.
Banning devices in schools isn't adapting to the new medium; it's just trying to preserve print-culture minds in an AI-native world.
Instead, we should be teaching kids how to think with these tools: interrogating algorithms, testing hallucinations, and treating AI as a sparring partner/oracle of Delphi rather than an answer machine.
This is absolutely the right approach by Winchester College. At school and university, access to all devices needs to be rationed if young people are to learn to read, think, calculate and write. Traditional cognitive skills are more not less important in the age of AI. thetimes.com/uk/education/ar…
33
See this doing the rounds. Yes they do say Mistral doesn't align with their climate goals but that is a footnote in a litany of other failures. The main being that it just isn't good enough (nor actually sovereign).
Berlin-based search engine Ecosia, used by government services, is dropping its French AI partner, Mistral, its CEO told POLITICO.
The company will now bet on open-source models, including those from China.
vist.ly/5mh5y
2
34
The problem of course is that if they are to be believed then superhuman AI will have already gamed out their game-outs.
Scoop: Top executives at Anthropic, OpenAI and other AI companies are privately gaming out scenarios for a public and political revolt after a catastrophic AI event…They assume one in ‘27…
axios.com/2026/10/09/ai-comp…
2
44
This feels wildly off the mark.
"AI's adoption will likely be even slower [than electricity], because companies have to reorganize around it."
1) Organisations aren't as hardcoded as physical places in 2026.
2) Workers in the 1800s weren't AI agents trained on the worlds knowledge.
And that's only one thing in here...
33
Elliot Leavy retweeted
Finland orders halt to work on two Google data centres bbc.in/3VwT5Uz
64
386
1,332
287,449
Welcome to the Stack War. Where the Cold War divided the world into First, Second and Third, AI is sorting it into the American stack, the Chinese stack and the rest.
Probably my favourite bit of the latest update to the MPI.
1
35
lol.
you couldn't have released this a day earlier and saved me the embarassment? @arthurmensch @IsForAt
2
58
I've been tinkering away at the Machinepower Index and in September I went deeper and audited the whole thing: more public data, strategy types that follow the evidence better, and a map of the two AI blocs taking shape.
Update notes: machinepower.substack.com/p/…
Index: live.machinepowerindex.org/
21