Fusion is getting the funding model every public programme should have: pay for milestones, not promises. Germany's SPRIND just opened €500 million through 2029 for fusion startups. Independent experts check each technical and financial milestone. Miss one and you leave the programme and lose the money. The last milestone is a design report for a demonstration plant. The US Department of Energy has run the same logic since 2023: a fixed payment only after a milestone is met, and the companies carry most of the cost. Commonwealth Fusion Systems said in August that SPARC was almost 80% complete. A target. A check. An exit. Imagine every subsidy in Europe written that way.

Google just showed the most inspection-minded idea in quantum computing. Instead of chasing a perfect qubit, you make a noisy state, test it hard, and throw away every flawed attempt. What survives is clean. They call it magic state cultivation. On the Willow chip, published in Science last week: errors cut about 40 times, and a record logical fidelity of 99.99%. Why it matters: by some estimates, making these states could eat up to 95% of a quantum computer's capacity. Cheaper magic means a smaller machine for the same job. Google is clear it still has to prove this at larger scale. In marine inspection we never trusted a perfect weld. We trusted a good test and the discipline to reject. Same principle, inside a chip.

A rocket that misses its target gets a post-mortem within days. A public program that misses its target gets next year's budget. That is the real gap between engineering and administration. Not intelligence. Feedback. Belgium spends 54.2% of its GDP through the state. France, 57.2%. Singapore runs a rich, safe, clean country at about 15.5%. Same century. Same technology. 3.5 times less state. Yes, Singapore funds pensions and health through mandatory savings accounts. That is the lesson. Design beats size. So here is how I would run a state if I treated it as an engineering problem. Musk's five steps, in order: 1. Question every requirement. Each rule carries a name, not a department. 2. Delete. If you never have to add anything back, you did not delete enough. 3. Simplify what survives. 4. Accelerate the cycle. 5. Automate last. Never automate a step that should not exist. Then one loop, forever: planned, actual, gap, correction. Published. This is where AI changes the game. Not by "governing". By measuring. Every program can publish its plan, its spending and its results. A second, independent AI can audit the first one, line by line, every week. "Who is accountable when the AI gets it wrong?" Fair question. Who is accountable today? With full traceability, every decision has a log, a version and an owner. That is more accountability than most ministries have ever had. Citizens set the rules and the budget envelope. The machine executes and reports in public. Any program that cannot show a measurable result gets an end date. The UAE appointed the world's first Minister of State for AI in October 2017 and adopted a national AI strategy to 2031 in 2019. It does not run its government by AI. It simply asked the question early. One self-taught citizen with an AI can now read a budget faster than a committee of hundreds. The question is no longer whether the state can be audited. It is who audits first.

En 1830, acheter les bougies pour une heure de lumière à 1 000 lumens coûtait environ trois heures de travail. Aujourd'hui, la même lumière coûte environ 0,07 seconde de travail. Trois heures, c'est 10 800 secondes. Le prix de la lumière, mesuré en temps de travail, a été divisé par près de 150 000. Aucun plan quinquennal n'a fait ça. Des inventeurs, des usines et la concurrence l'ont fait, essai après essai. J'y pense chaque fois qu'on me dit que l'IA coûte trop cher. La bonne question n'est pas le prix d'aujourd'hui. C'est la pente.

Humanity runs on about 19 terawatts. The Sun puts out about 380 trillion trillion watts. The gap is roughly 20 trillion to one. Closing it is what "Kardashev II" means: a civilisation that uses the energy of its star. Yesterday Elon Musk listed the only things he thinks scale that far: superintelligence, robots, solar power and mass to space. And if you have the first two, the rest follows. I agree with the order. Intelligence and hands were always the scarce inputs. Make both cheap, and energy and launch become build problems instead of dreams. Inside a small company the same order works, just smaller. Make the thinking and the doing cheap first. The rest follows.

4,000 procedures. Gone. That is what the UAE reported in June 2025 for the first phase of its Zero Government Bureaucracy programme. More than 4,000 unnecessary procedures eliminated. 1,600 redundant requirements removed. Service times cut by over 70%. About 12 million hours a year handed back to people and companies. Phase two went after digital bureaucracy. One work-permit service went from 72 data fields to 37. Delete before you optimise. That is step 2 of Musk's five-step method, run by a government. What I like is not the slogan. Someone counted. A deletion you can measure is a policy. A simplification you cannot measure is a speech. How many procedures did your country delete last year? Does anyone even keep the number?

France has the resources. 57.2% of GDP goes through the state, 3.7x Singapore's 15.5%. The problem was never money. It's design: no rule with an owner, no program with an end date. Engineer the state and a week becomes credible. xitter.cf/nayibbukele/status/211…