Anglofuturism

Part two of our conversation with Will Orr-Ewing gets into the harder questions: whether a genuinely meritocratic elite is more dangerous than an aristocratic one, why AI tutoring has solved the wrong problem, and what it would take to build an Odyssean education for Britain’s most talented kids.
Tom, Calum, and Will discuss:
* The internet should have produced a generation of Einsteins — it didn’t: Eric Hoel’s provocation that the most naked conclusion you can draw from the internet, and now AI, is that the constraint was never information availability. The knowledge was always there. We’ve done something bad to intrinsic motivation. “Where are all the people who used the internet to teach themselves untold knowledge?”
* Why AI tutoring has solved the wrong problem: Alpha School puts children in pods on the 35th floor of a New York skyscraper, not allowed to communicate, staring at screens. Will’s friend visited and saw four tantrums in a single school trip. The problem isn’t personalisation — it’s that children don’t need education adapted to their interests. They need their interests adapted to what’s worth learning. And AI cannot do the one thing that actually works: be someone a child wants to become.
* The meritocracy trap: A genuinely meritocratic elite is a terrifying thing. They owe nothing to anyone because they earned everything themselves. Whereas the aristocrat could never quite believe he deserved his position — it was an accident of birth — and so noblesse oblige followed naturally. “You look at the winners of the last 20 or 30 years. They just don’t seem to have a sense of obligation to their country.”
* The Odyssean curriculum — Britain as the school of the world: Cummings’ essay argued England could be what Athens was to Greece — a model for how to educate statesmen and scientists. Will wants an Odyssean version of the King’s Maths School from age 14: Thucydides, Lee Kuan Yew, applied geopolitics. Cohort effects like the Brit School at the Grammys. Currently the maths olympiads have barely 600-700 entries a year. “Our future disproportionately relies on those people. And at the moment their track leads to being a quant at a hedge fund.”
* Elite kids as asset managers of their own human capital: Daniel Markovitz on how the most ambitious families in the world — Will has offices in Singapore, Hong Kong, London — are depleting their children through constant striver credentialism. Nonverbal reasoning tests that you forget the moment you’re through them. “If it was Dostoevsky, at least it might stay with you. But most of these competitive entrance exams have no enduring value whatsoever beyond your LinkedIn trajectory.”
* What Will actually wants for his children: Walking through Parliament and knowing every statesman on the wall. Walking through the countryside and knowing every tree, every bird. “Education properly done is a vitalising force which enchants your everyday perception.” And one other thing: if they’re in a room of a thousand people and 999 say sign the document, the moral courage to say no.
Plus: Rory Stewart’s dad recreating Waterloo in Hyde Park before school, the Anglofuturist Great Hedgerow of Britain as a children’s internet firewall, Korean tutoring centres prohibited after 10pm, and whether Singapore has started workshopping “thinking outside the box” with an actual drawn box.


This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.anglofuturism.co/subscribe

Will Orr-Ewing has spent 20 years tutoring and founded Keystone Tutors, but he’s not here to tell you to hire a maths tutor for your nine-year-old. His argument is bigger: that Britain once had a culture of self-directed intellectual growth that state schooling quietly strangled, that the billion-pound tutoring industry is almost entirely pointed at the wrong goals, and that the GCSE system is simultaneously boring the top of the cognitive distribution and failing the bottom.
Tom and Calum receive him in the somewhat dusty schoolroom of the King Charles III Space Station to design an Anglofuturist curriculum—and debate whether the state can ever do what a parent, a tutor, or a good book can.
Tom, Calum, and Will discuss:
* Why tutoring is a superpower pointed at mediocre ends: “You’ve got this massive potential for intellectual expansion, but directed at very menial, mediocre ends.” The billion-pound industry is almost entirely Kumon-style drilling or GCSE cramming. The mimetic relationship between tutor and student—where the neophyte absorbs not just knowledge but how someone thinks—is almost entirely wasted on exam prep.
* The autodidactic culture that state schooling killed: Before the 1870 Education Act, elite education meant acres of childhood time for reading, with tutors as a clinic to check progress rather than the engine of learning itself. “All education is self-education,” as Charlotte Mason put it. The state provided for the bottom but quietly smothered that instinct everywhere else.
* GCSEs are failing everyone except the middling: Thirty percent fail maths and English GCSE every single year. The top of the distribution is bored stiff. “It’s only the middle runners who are really being served.” Schools are so incentivised to chase results that any choice between intellectual stretch and hammering assessment objective three goes the same way.
* The case for releasing kids at fourteen: The bottom thirty percent for whom the credentialist conveyor belt—GCSEs, university, graduate scheme—is “clearly so unenticing.” A more apprentice-based model, local relationships with employers, learning a trade. Michael Faraday was a bookbinder’s apprentice for seven years. A lot of fourteen-year-olds would rather be on an Isambard factory floor than in another PowerPoint-driven lesson—if the smartphone weren’t in their pocket.
* The state cannot replace parental culture: “The real problem is that the state cannot replace the role of a genuine parental culture.” Any attempt to enforce it through the curriculum cheapens it. The dirigiste continental model—school as nation-building—turns what was once emergent into a bureaucratic goal liable to be rewritten by a single pen. And yet: do we trust modern parents to deliver? “I’m not sure I do.”
* Schools as the last mile of the welfare state: Teaching children to use the loo. Brushing teeth. Breakfast clubs. “Whenever there’s an issue we decide as a society that we care about—the environment, AI literacy, financial literacy—it gets shoved into the curriculum, further bloating it and further undermining the chances of delivering something excellent.”
* The Anglofuturist village school prospectus: Gowns and mortarboards. Blackboards. History running from Æthelstan rather than Rosa Parks. Drone-building classes. A wall between the boys’ and girls’ houses patrolled on a mathematically complex schedule—crack the algorithm, and what awaits you is left as an exercise for the reader.
Plus: why Æthelstan would be confined to a cartoon on a Twinkl worksheet even if teachers wanted him, the left-wing case for aristocratic tuition, education savings accounts in half of American states, and whether sourdough is woke.


This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.anglofuturism.co/subscribe

In part one, we explored why drug development costs are exploding and how better software could fix it. In part two, we get practical: what’s actually stopping Britain from becoming a biotech superpower, and what would it take to get there?
Meri pulls no punches. The single hardest thing about building Lindus Health in the UK? Three-month notice periods. Want to staff up for new trials? Wait three months for people to work out their notice—during which they’re not exactly doing their best work. “It’s incredibly ineffective. It acts as a transfer from the most productive companies to the less productive companies people are resigning from.” Meanwhile, US contracts have no notice period or a couple of weeks max.
But notice periods are just the start. The real bottleneck is that Britain produces excellent early-stage research but can’t capture the value because we’ve made ourselves an unattractive market for drug sales. NICE’s role has become “get the lowest price possible, even if that means greatly delaying when the drug is distributed in the UK.” We’ll spend five years negotiating a thousand pounds off a course of treatment while people literally die. The solution? Turn the NHS into a pharma company—have it fund and run trials like the RECOVERY trial that discovered dexamethasone, then earn royalties by selling the drugs to America.
From ethics committees run by religious volunteers who delay STI trials to promote abstinence, to why Brexit was actually good for medical devices (FDA approval now automatically carries over to UK), to the limits of in-silico trials and why randomised control trials are “literally magic,” Meri lays out a vision for fully automated luxury NHS—and explains why everything comes down to clinical trials, even in the age of AI.
Tom, Calum, and Meri discuss:
* Why Meri’s company had to go transatlantic: “We haven’t moved to the US—we’re transatlantic. About 150 people, half still in UK. But look, I’m not going to deny there are strong forces pulling us to the US.” Not capital availability—European investors funded them to Series B. It’s the market. “Markets aren’t big enough in Europe to sustain global category dominant companies. If you want to build category defining companies in the UK, you need to grow the economy.”,
* Three-month notice periods are killing British startups: “The single hardest thing about building the company so far in the UK has been notice periods. We’ve won new trials, need to staff up, hire good people—takes time. Then they have three months between resigning and joining us. It’s incredibly ineffective because once you’ve resigned, you’re not doing your best work.” US contracts: no notice period or couple weeks. “Even a couple weeks is enough to fully hand over even a senior productive person’s work.”,
* What Lindus Health actually does: Design overall study, find sites, train them, oversee operations through software that integrates with health records and labs. Monitor for errors and patient safety risks in real time. For home-based trials like ME/CFS: “We employ nurses directly to visit patients in their home or have video calls. We do pretty much everything.”,
* Adaptive trials that analyze data in real time: “Clinical trials today are very waterfall. Design, run, analyze months after it’s wrapped up. Our software runs every trial adaptively. We don’t know how many patients we’ll enroll or what ratio between control and treatment. Software automatically randomises patients in a way that boosts statistical power and stops enrolling as soon as we’ve enrolled enough to show statistical effect.” Not p-hacking—stays blinded,
* Testing multiple variations in parallel: “Should be testing multiple in parallel. One control arm of 100 people, indeterminate number of arms with slight variation of dose or patient population. For the same time and massive cost saving, get way richer data.” Already doing this today,
* Why in-silico trials are limited: “RCTs are literally magic. By randomizing participants fairly, you control for all possible variables without needing to know what they are. To run effective in-silico experiments, you need to know what all possible variables are, which is essentially impossible because humans are incredibly complex.” Where they work: late-phase cancer (unethical to give placebo) and psychedelics (you immediately know if you got ketamine),
* Brexit was actually good for medical devices: “If you get FDA approval for a medical device, you automatically get approval in UK—been a big triumph post-Brexit. What would be amazing is to have it both ways.” For drugs, you still need slightly varying requirements for each country but one expensive phase three gets approval in Europe, Japan, US, South Korea,
* Ethics committees run by unhinged volunteers: “Someone delayed phase two oncology trial—so people were going to die—because they felt the font was too small in documents. Delays by at least four weeks because the committee only meets every four weeks.” One person delayed STI test trial because of religious conviction, insisted on promoting abstinence,
* Just pay for private ethics committees: “In US you can pay private regulated company to convene ethics committee. Costs five or ten grand but we get quick good feedback and can start in a week. That’s a no-brainer—same centralised system but pay the people, implement rigorous standards, make it self-funding.”,
* The COVID trials that worked: Recovery trial—Martin Landray ran very fast pragmatic trial testing different COVID treatments. Discovered dexamethasone was effective at reducing mortality. “Extremely cheap in drug trial terms.” Their VP of clinical operations was key person behind panoramic and principle trials, both fully remote. “By really tight integration with health system, you can run trials so much faster and cheaper in a way that’s not possible unless you are the health system.”,
* Turn the NHS into a pharma company: “Have NHS run trials for free or very low cost like RECOVERY. In return they own a share of the drug. We’ve run phases 1-3 on NHS very quickly—now we’re the distributor or we sell license to pharma and earn significant royalty. British patients get access sooner and it would be incredibly profitable because you run these trials so much cheaper than on US healthcare systems.” Would require fundamentally re-architecting NHS around for-profit model,
* Why speed matters more than people think: “Because of how patents work in life sciences, every day that ticks by is literally on average worth hundreds of thousands for the average drug. That’s less revenue you could be earning before the patent cliff when drug goes off patent and becomes generic.” Speed should be incredibly important—and they reinvest that revenue into fundamental R&D,
* The vision for 50 years from now: “If we can crack opening this bottleneck—safely test 10x, 100x as many iterations of potential drugs at scale—you inevitably get healthcare bioabundance. This has to happen to cure cancer, cure Alzheimer’s, live to 200. Everything comes down to clinical trials. Until AGI can completely simulate the human body, you literally cannot objectively claim you’ve cured cancer until you’ve tested it in enough humans.”,
Plus: Why lipids massaged into mouse hair could cure Tom’s Norwood 2, the meeting rooms named after James Lind’s original trial arms (cider, seawater, oranges, lemons, barley water, garlic paste), and why they randomised people onto different drinks at their early parties.
Your clinical trial success depends on notice periods—who knew?


This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.anglofuturism.co/subscribe

This is a free preview of a paid episode. To hear more, visit www.anglofuturism.co

It’s Valentine’s Day morning and Calum woke up with Rupert Lowe promising to restore Britain. Tom made bacon sandwiches and tea with plenty of sugar. And Britain’s vertical launch dreams just died—Orbex, the country’s great hope for homegrown rockets, has collapsed into administration. Is this a tragedy or were they building the wrong rockets all along?
What follows is a sprawling argument about whether Britain should mourn or celebrate, why the government won’t fund proper space ambitions, and the deeper aesthetic war underlying every political debate in this country. From Jim Ratcliffe’s “colonisation” comments triggering the PM to demand an apology, to the question of whether HCBGs or keffiyeh-wearing Oxfam shoppers represent the real Britain, Tom and Calum diagnose why we can’t have nice things—and what it would take to build an O’Neill cylinder with cricket fields anyway.
Tom and Calum discuss:
* The Orbex collapse as Britain’s space 9/11: The vertical launch company went into administration after a Franco-German takeover fell through. Tom mourns the loss of Union Jack rockets. Calum says “they were building the wrong rockets”—small satellites when SpaceX’s super heavy lift has made that model obsolete. “We’re doing this weird combination of all space in industry, very little government funding, but we want the totemic sexy capabilities. We’re not providing a market for them.”,
* Britain’s actually brilliant space sector: Space Forge with their 1,000°C furnace and Pridwen heat shield named after Arthur’s shield. Surrey Satellite Technology Limited pioneering shoebox-sized satellites. Astroscale doing “space MOTs”—fixing and removing orbital debris. “We do have a pretty cool space sector in terms of the small stuff, the space engineering frontier.”,
* The milestone payment model SpaceX used: “You offer fixed amounts of money as milestones. If you hit the milestones you release more.” Tom Kalil’s Renaissance Philanthropy approach. “If you put up money for competitions you only have to pay out if you get the capability.” Far better than cost-plus contracts that create infinite money pumps and overruns,
* The regulatory sandbox is actually good: Companies working on new space tech can “send someone to sit in a room with someone from DSIT and come up with regulation in real time.” If you want to test nuclear propulsion in space, “the cold hand of DSIT reaches out even that far. It will gently tickle you instead of totally throttling you.”,
* The mythic quality Britain’s missing: Lord Kempsell asked what the plan was to get an Englishman on Mars. No answer. “I think it’s the mark of a healthy country to have that kind of ambition. I think it’s good to foster the ambition of young men who might wish to die defending the British settlement on Olympus Mons.”,
* The real aesthetic war in Britain: Not HCBGs vs reformers. It’s “Green Party style—privilege is bad, keffiyeh as your style statement, women with quite short-cropped hair, big boots, Doc Martens with Superman socks. A kind of lower-middle-class earnest, very morally fierce Britain of suburban middle towns.” In Cornwall: coastal towns are “Joules, Jack Wills, Helly Hansen, HCBG Central.” Inland towns: “two vegan restaurants and an occult bookshop.”,
* Jim Ratcliffe and elite defection: Said Britain has been “colonised by immigrants.” PM demanded apology, Number 10 welcomed it when he gave soft apology. Tom’s friend on football group chat: “plainly racist.” Tom: “I don’t think it is. Strong meat linguistically, but not plain racism.” Government wasting time on words instead of integration issues,
* Why the PM shut it down so fast: Not strictly semantically accurate but “there is something in it—whole areas have changed, people staying, sending remittances home, organized crime. You could say there’s some truth to the word colonisation. The fact there’s some truth to it is why the PM has been so quick to shut it down.”,
* The progressive theory of speech codes: “If you punish people hard enough for breaking the speech code, the problem will go away. Because there was no problem anyway. The problem was the working class getting false consciousness because of elites like Ratcliffe.” So you punish Ratcliffe at the source—despite him being a tax exile which doesn’t help his public image,
Plus: Tom’s 10pm tatty scone gammon eggs Benedict, why Calum thought he’d be grooming talent at Civic Future, the milkman arriving at KC3, fake smoke allegations at British rocket companies, and whether frame-mogging Chinese astronauts requires large bums like skeleton bobsledders.

The pharmaceutical industry has a dirty secret: it takes $2 billion and a decade to approve the average drug, and these numbers are getting exponentially worse. While computing power doubles every few years, drug development costs double every decade—a phenomenon called Eroom’s Law (Moore’s Law backwards).
Lindus Health was founded to fix this crisis. Named after James Lind, the Royal Navy surgeon who ran the first randomized controlled trial in 1747 (discovering that citrus prevents scurvy and accidentally creating the Sicilian Mafia in the process), the London-based company is slashing clinical trial costs and timelines through better software, smarter processes, and a willingness to actually keep up with FDA guidance—which, remarkably, the industry ignores.
In this first part of our conversation, we explore why pharmaceutical shelves are lined with miracle drugs gathering dust, how the NHS simultaneously possesses world-class health data while being catastrophically bad at purchasing new treatments, and what Britain could gain by becoming the world’s biotech testing ground.
Tom, Calum, and Meri discuss:
* Why drug development costs are doubling every decade: Eroom’s Law means $2 billion and 10-12 years per drug on average. “A tech bro would say ‘it one shot me’ right? How have we got this incredibly important industry getting exponentially less efficient when all the inputs—genome sequencing, compute—are getting exponentially more efficient?” The vast majority of costs are in phase 1-3 clinical trials,
* The COVID vaccine trials were archaic: Meri volunteered and “it was like stepping back 30 years.” He had to download Microsoft Edge because the signup website didn’t have an SSL certificate. “That sounds trivial and silly, but that probably puts off at least half of potential volunteers, which makes it twice as long to enroll and potentially twice as expensive.”,
* Pharma shelves are lined with miracle drugs gathering dust: “You would be shocked. There are just umpteen compounds sitting on shelves gathering dust.” Often shelved for ridiculous reasons: “This was a pet project of this guy who got fired and no one else wants to touch it.” Or outdated NPV thresholds. Because trials are so expensive, it’s not worth their time,
* The regulations are surprisingly permissive: “This will sound controversial but I think the regulations have an appropriate level of risk modulation. You can literally go on the FDA’s website and see briefing documents where they are admonishing pharma for not being innovative enough. What other industry is the regulator trying to force private companies to be more innovative?” Most barriers are self-imposed,
* James Lind and the Sicilian Mafia: In 1747, Lind ran the first RCT to cure scurvy—up to 50% of sailors on long voyages just died. Six treatment arms, oranges and lemons won. “One of the key innovations that powered the British Empire.” The demand for citrus was so great the Royal Navy went to Sicily, and “the Sicilian Mafia formed as a collective bargaining organization to help producers get a fair price.”,
* The low-hanging fruit argument is cope: “Most people would say ‘oh well maybe we’ve discovered all the early targets and all that’s left is really hard to drug.’ That just seemed like terrible cope. 30-40 years ago we discovered medicines by zapping them into mice randomly. Now we’ve sequenced the human genome.”,
* Britain has incredible advantages it’s squandering: The NHS has “probably the best health data set in the world. Completely longitudinal cradle to grave, all one system, records coded the same way.” UK Biobank is world-class. “There’s a lot of early phase research that originates in the UK. But when you’re running later trials, you want a good early adopter market. That unfortunately is not the UK.”,
* The NHS purchasing problem: NICE has decided its role is to get the lowest price possible “even at the expense of waiting five years to acquire a drug that could be life-saving. We’ll spend five years negotiating a thousand pounds off a course of treatment and you think, is that worth it? People are literally dying who could have not died.”,
* The dream scenario for Britain: “The NHS will fund the entire clinical trial and in return the drug will be free on the NHS. Maybe the NHS earns money off royalties of sales in other markets. That would be incredibly powerful, incredibly accretive to the British economy, but it would require political will.” If everyone’s going to worship the NHS like a deity, at least make it productive,
* GPs are secretly based: They’re “basically private companies and thus much more flexible and fast and easy to work with” than NHS hospitals. Lind runs many trials through GP surgeries and patients’ homes to avoid hospital bureaucracy,
* The ME/CFS trial: Running a trial for chronic fatigue syndrome with a German pharma company entirely remotely because “the sickest patients are literally bed-bound.” Using a drug already approved elsewhere. “I don’t care how the disease mechanistically works. I just care that we can run a proper experiment. If it works, I kind of don’t care how it has an effect as long as it works.” Testing beats theory,
* Why the industry won’t innovate: “Incredible inertia driven ultimately by pharma having huge regulatory barriers to entry and thus very little competition and thus little pressure to innovate.” COVID vaccines succeeded because there was “for once, intense competition.” The problem isn’t that regulations are too strict—it’s that nobody bothers to follow guidance that would make things faster,
* What Lindus Health actually does: Makes clinical trials faster and cheaper through better tech and processes. Uses AI to generate higher quality trial documents, quality control protocols, find patients more efficiently. $80 million raised, majority of trials now in US because “healthcare market is dominated by the US.” Over half of clients’ trials are American,
Plus: The hellish anti-snoring device, why thalidomide broke our risk tolerance, how decentralized trials work, the bitter lesson of machine learning applied to pharma, and why Meri thinks Britain could create the next Novo Nordisk if we just got our act together.
Part two coming soon.


This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.anglofuturism.co/subscribe