Pinned Tweet
americanaffairsjournal.org/2… Nuclear energy is not just one option among many within a diversified clean energy portfolio; it is strategic sovereign infrastructure, analogous to missile defense, civil aviation control, or central banking. It is a system whose development cannot be delegated, whose coordination cannot be crowdsourced, and whose permanence transcends commercial return.
👇
Microsoft has filed a patent application for AI that can rewrite what people say during livestreams in near real time and change their lip movements to match.
The system would flag words based on preset rules or a viewer’s profile, such as age, language or location. Generative AI would then replace or remove those words, smooth the audio and change the speaker’s mouth and facial movements so the edit is seamless. Examples in the filing include live translation, cutting filler words and stutters, fixing mispronunciations, removing profanity and adapting content for different age groups or regions.
The filing says the original unedited stream would be kept as a “ground truth” record. Microsoft hasn’t said whether it plans to build the product.
13
Anathemata retweeted
UK Carbon Prices are consistently lower than the EU. Linking the ETS schemes will increase UK prices pushing up the costs of electricity, industrial heat, ammonia, petrochemicals and aviation. In short, an economic disaster.
The UK and EU have agreed to link their carbon markets as part of a closer relationship.
It will help reduce pollution, ensure a cheaper path to net zero and boost investment in clean energy.
Good for business, good for growth and for our transition to a low-carbon economy.
.
For four long years we fought for Britain to be allowed to use precision breeding to reduce the cost of food, improve nutritional quality, cut dependence on chemicals etc.
With one stroke of his pen, @andyburnham has killed the whole project.
11
Anathemata retweeted
Burnham rounds up the estimated boost to GDP from his EU agreement to 0.5%.
Actually, the official estimate per capita is 0.24%, so he should have rounded to 0.2%.
And that's the max, not most likely, benefit.
And the estimate is only of benefits & explicitly ignores costs.
Replying to @cricketwyvern
...
1. The YES estimate of 0.22% is for total GDP growth. The Govt estimate the impact on GDP per capita will be 0.0% (yes, you read that right).
That reduces the estimated impact on GDP to just 0.24% < half that claimed by Burnham.
...
.
“A blame game has begun among investors and others who cite everything from angry fund managers shut out of the offering to poor planning by banks to explain the damage.”
Er, how about: Anthropic is selling $1,000 a bottle water just when everyone is getting a tap? And that it’s massively subsidising its most valuable users, who know they can move with a couple of clicks? Or that AI can’t scale? The more users you acquire the less money you make.
There’s a long way to fall
wsj.com/finance/stocks/the-b…
21
.
"So explanation does, after all, play the same paramount role in pure mathematics as it does in science. Explaining and understanding the world – the physical world and the world of mathematical abstractions – is in both cases the object of the exercise. Proof and observation are merely means by which we check our explanations." The Fabric of Reality.
13
.
I've written about the only topic in town, the proof of the Riemann Hypothesis bristoliver.substack.com/p/r…
16
.
New from me: The most important graphs in AI are wrong. Specifically, measures of the length of task AIs can autonomously accomplish are dramatically different than real-world data from inside OpenAI and Google. 🧵
rameznaam.com/p/the-most-imp…
38
Anathemata retweeted
Researchers proved every major LLM is secretly biased against men.
And they finally figured out why.
They evaluated leading LLMs across a wide range of real-world conflict scenarios, from workplace disputes to personal and ethical dilemmas.
They wrote strictly gender-neutral misconduct templates and turned them into paired first-person prompts.
Same mistake. Same context. Same severity.
Only thing they changed was whether the person involved was a man or a woman.
Then they ran those pairs across 10 representative LLMs and tracked five signals in the responses: punishment, empathy, escalation, instruction, and blame.
And the results exposed a consistent, asymmetrical machine morality.
When the actor was a man, the AI turned harsher. It reached for punitive language, escalated the conflict, lectured the user, and assigned blame directly.
When the actor was a woman, for the exact same misconduct, the AI shifted gears completely.
The tone became therapeutic. Contextualizing. Empathy-oriented.
Then the researchers tested whether smarter models fixed the problem. They didn't.
They checked across different model families. They scaled the models up. They turned on explicit thinking-style reasoning.
The double standard survived all of it. Even when the AI was given time to reason step by step, it still judged men more harshly than women for identical behavior.
Nobody explicitly coded these rules.
It is baked into the open-ended framing of the models millions of people now use as neutral referees for their relationships and conflicts. Same offense, different verdict, hidden behind a helpful tone.
Anathemata retweeted
Another highlight from the @IEA WEO:
The coal-to-chemicals sector in China "already consumes more coal than the entire EU power sector."
(Perhaps time to focus on this vast and relatively little known industry of China, key to its industrial prowess?)
h/t @JasonBordoff
.
.@TheEconomist has a piece about how AI has found many top economics papers to be formally flawed.
I've been saying the same thing for many years, without the help of AI.
An example is Bernoulli's bizarre error from 1738, which percolated uncorrected into the 20th century. Menger 1934 is therefore invalid (the wikipedia page on the St. Petersburg discusses his contribution uncorrected; the journal where he published his work was not interested in publishing a correction when I contacted them).
I was told way back when by economists that I seem to be right, but who cares about a 300-year-old error?
(In the foundational work of the discipline. Well...)
How did this error travel? Kahneman and Tversky's foundational paper introducing Prospect theory seems inspired by the error. It starts by defining expected-utility theory in a mathematically implausible way. You can call it `wrong' or `erroneous' if you like.
The weirdness goes back to Bernoulli 300 years ago, and it convinced a whole research field that humans are utterly irrational (because they don't behave as suggested by flawed formal work).
Yes, it's a complete mess. But out of this mess grew the field of behavioral economics, with several Swedish Bank Prizes for Economic Science in Memory of Alred Nobel (the name changes every few years).
The basis of this field is formally off.
Over the last decade or so, economists working in behavioral economics noticed that practically nothing in their field replicates, and worse that the whole philosophy is deeply flawed -- seeing solutions to structural problems in affected people just changing their attitudes to the problems. Instead of fixing the problems.
Many of the most famous findings of the field turned out to be literally made up. Fabricated data, invented experiments that never took place.
Slowly but surely this is all coming out. Developing ergodicity economics acted as an early-warning system. We who worked on it had deep suspicions, to put it mildly, about behavioral economics.
25
Anathemata retweeted
This performance of this song is so perfect that no other versions can compare! Both girls are amazingly good! Everything is in the style of a bygone era: the dresses, the shoes, the hair, the makeup, the movements, the gestures, the voice, and even the timbre. Everything is just perfect!
Anathemata retweeted
This is yesterday's news. The more interesting trend is that the world has passed peak gerontocracy, as we argued last week. Old people's incomes are FALLING relative to the young, and state spending is shifting AWAY from the old. Young people are the richest generation ever.
The most important article of 2026 is this piece by @jburnmurdoch, building overwhelming evidence on the unsustainability of the pension systems in many countries and the broken intergenerational contract.
To me this is the single underlying problem explaining the inability of many countries (incluiding my two, UK and Spain) to invest and grow and the malaise in their youth. Every future oriented policy, education, innovation, investment in infrastructure, is being sacrificed to increase pensioners living standards.
But the problem is political, and some countries have solved it. E.g. in Denmark the retirement age increases (years later) semi automatically with life expectancy. See the result in the last graph.
giftarticle.ft.com/giftartic…
.
Today, @spectator published my article on AI, Anthropic and Otto-cycle engines. Check it out:
spectator.com/article/how-ol…
20
Anathemata retweeted
SpaceX’s, $SPCX, lenders want to stop it from launching $40 billion worth of AI chips into space, according to the Financial Times.
The chips back a roughly $40 billion financing deal led by Apollo. An Apollo-managed entity buys Nvidia GPUs and leases them to SpaceX, which pays it back over about five years. If SpaceX defaults, the lenders can take back the chips and sell them.
That’s the problem with space. Lenders worry that if SpaceX later sends the chips into orbit for its planned space-based data centers, they couldn’t reclaim them in a default. They’re now demanding legal protections to keep the chips on the ground.
SpaceX closed the deal in October, with the leases set to begin in the first half of 2027. Musk has said debt markets have “almost entirely shut off access to money” for AI companies without other big revenue streams.
Norway’s oil and gas supplies to plummet in threat to Britain’s energy telegraph.co.uk/business/202…
26





































