Britain has a habit it cannot break. We invent the future, then we sell it to people who understand its value better than we do. This week a select committee told ministers to remove Palantir from the NHS, and the company's British boss spent the days that followed defending the contract like the politician he is. In this issue, David Richards MBE asks who really owns the most valuable dataset in the world, and whether we are about to repeat with our health records the mistake we made with DeepMind.

In this episode:

The £10 billion goldmine. Why Palantir hired a Westminster networker rather than an engineer, the holes in Louis Mosley's Telegraph defence, the Nick Clegg parallel, and the National Grid precedent for ripping foreign technology out of critical infrastructure. Plus the close that ties it together: how Britain built DeepMind, sold it to Google for around £400m, and handed Silicon Valley the very capability it feared.

• The 17-year itch. David's Digital Forge conversation with Richard Stubbs of Health Innovation Yorkshire and Humber on why proven NHS technology takes an average of seventeen years to reach patients, why there are "more pilots than British Airways," and the safety question nobody dares ask.

• Would AI have locked us down? David's Yorkshire Post column on the Imperial model, the statistics watchdog's rebuke of the Covid Inquiry, and what AI-assisted models would actually tell a minister to do today.

• The layoff tracker, Week 23: US tech's heaviest month of cuts in nearly two years, and the mechanics of AI laundering.

• Read the full issue, with every source and link, free at Https://TheSundaySignal.ai

Sources this week include: the Science, Innovation and Technology Committee report; Louis Mosley's op-ed in the Telegraph; reporting from Reuters, the Financial Times and Challenger, Gray and Christmas; Ernst and Young's valuation of NHS data; and trial exhibits from Musk v OpenAI and The DigitalForge Podcast

Whatever you believe about AI, check the date on it. In this field, out of date means days, not years.

This week three things landed together: a Sequoia keynote arguing this is a revolution in computation rather than communication, a Y Combinator chief executive shipping like a team of twenty, and a banking boss caught describing 7,800 job cuts as "replacing lower-value human capital."

David Richards MBE connects them, then turns to the economics (one firm reportedly spent half a billion dollars on Claude in a month) and the human cost recorded in the weekly Sunday Signal Layoff Tracker: 472,150 cuts so far this year, 169,430 of them in tech.

In this episode:

• Why Pat Grady says "the floor keeps moving underfoot"
• Garry Tan, gstack, and one builder shipping like twenty
• The survival playbook for anyone whose job is a task
• The 725 billion dollar capex wager and the dot-com parallel
• Standard Chartered, ClickUp, and the algorithm in Thatcher's coat

Read the full written edition and subscribe free at thesundaysignal.ai. A new issue every Sunday.

The AI energy debate has been built on a single assumption. That hyperscalers will buy raw electricity at any premium because they have no other choice. A British company in Oxford has the receipts to prove that assumption is wrong.

This week, the full interview with Dr Xianxin Guo, CEO of Lumai, the only company in the world running billion-parameter LLMs on optical compute. Ninety per cent less energy. Cost per token that rewrites the economics of inference. And the procurement-not-subsidy demand he is making of the British state.

Plus the layoff tracker hits another 18,800 this week, led by a 170-year-old British bank telling investors it is replacing "lower-value human capital" with AI, the most candid C-suite line of 2026 so far.

Three stories. Photons. Power. Britain.

The full newsletter and interview at thesundaysignal.ai. Read this week's issue here: https://TheSundaySignal.ai

The professional CEO era may be ending.

For decades, the moment a technical founder built something valuable, investors, universities and boards reached for the same answer: install adult supervision. Bring in the polished operator. Add process. Build the reporting layer. Professionalise the company.

This week’s Sunday Signal argues that model is breaking.

At a Sheffield spin-out event, in the OpenAI courtroom battle, and inside the rise of AI-powered operating systems for founders, the same truth is becoming visible: much of what professional management claimed as leadership was really administrative machinery.

AI is now absorbing that machinery.

Meetings, reports, follow-ups, investor updates, hiring admin, board packs, market intelligence and financial narratives can increasingly be handled by AI agents. That does not replace judgement, taste, courage or accountability. It makes those qualities more important.

In this episode, David Richards examines why “adult supervision” was often the problem, why technical founders are about to regain leverage, and why Founder OS may become the new operating model for founder-led companies.

Founder-led. AI-powered. Professional management was the old operating system.

Full newsletter, complete Founder OS playbooks and the full layoff tracker at: https://TheSundaySignal.ai

Issue #53 of The Sunday Signal. Sunday 10 May 2026.

Healthcare is the most regulated, slowest-moving and institutionally defended industry in the modern economy. This week, all three of those defences cracked in the same direction.

Three stories.

One. The FDA has announced AI will monitor drug trials in real time. AstraZeneca and Amgen are already named as proof-of-concept partners. The Commissioner publicly concedes that 40 per cent of clinical trial time is "dead time" AI will dissolve. This is not an efficiency programme. It is a regime change. Britain's MHRA has nothing comparable live in 2026.

Two. The patient builds the model. Two years ago I sat in an A&E corridor with a family member showing suspected stroke symptoms, using AI on my phone to think through differential diagnoses. This year I have built a personal AI health model on my own data: Whoop biometrics, private blood panels, every NHS record I could obtain. GPs in both the US and the UK have missed things. They are human. The system that asks them to evaluate a multi-decade health history in eight minutes is the problem. AI does not have an eight-minute limit.

Three. The Fifth Wave. The single greatest achievement in human history is the doubling of life expectancy from 32 years in 1900 to around 73 today. Four waves of breakthrough got us here. The curve has been flat for two decades. The case for AI as the fifth wave. Why Silicon Valley's 150-year claim is wrong on timeline, but the squaring of the mortality curve is the real, achievable prize.

Plus the weekly Tech and AI Layoff Tracker. Cloudflare's first mass cuts in 16 years. Coinbase, Freshworks, Upwork and BILL Holdings. Microsoft's 900-million-dollar voluntary buyout programme. The polite fiction of "AI-assisted workforces" is over.

Read the full newsletter at thesundaysignal.ai.

Musk on the stand. LeCun staking a billion. And every adviser, founder and junior in Britain learning what happens when they hand theirs to a machine.

This week's issue opens in a federal courthouse in Oakland, where Elon Musk has spent seven hours under oath arguing that the men he funded broke a charitable trust to build the most valuable AI lab on earth. The most damaging evidence is not from Sam Altman. It is a notebook page from Greg Brockman that calls the nonprofit pledge a lie in plain English, written months after he had assured Musk it would hold. We walk through the Larry Page kitchen conversation that started OpenAI, the three-phase narrative that may sink Musk's own statute-of-limitations argument, and the moment William Savitt got Musk to admit, on the stand, that he had once told Tesla analysts he was building an enormous AI-enabled robot army.

We then turn to a Yorkshire negotiation last month, in which a senior professional handed an advisory board offer to a chatbot, asked for guidance, and lost the role inside twenty-four hours. The lesson, expanded from this week's Yorkshire Post column, is that AI is a multiplier. It amplifies expertise where there is a floor underneath. It exposes confusion where there is not. We offer five practical rules for using AI without becoming the person being quietly marked down.

We close with Yann LeCun and AMI Labs, the billion-dollar seed round built on the bet that the entire industry's foundational judgment has been wrong for five years. Plus the Week 19 layoff tracker, where Microsoft's first such programme in fifty-one years and Meta's eight thousand cuts pushed the year-to-date total past two hundred and seventy thousand.

Read the full issue, with charts and source notes, at thesundaysignal.ai.

For the first time in its 51-year history, Microsoft has offered voluntary retirement to thousands of its veterans. On Polymarket, real money is wagering at 90 per cent that 2026 will finish with more US tech layoffs than 2025. And in San Francisco, a 20-year-old from Texas threw a Molotov cocktail at the front gate of Sam Altman.

These are not three stories. They are the same story told in three registers: the boardroom, the market, and the street. All three are converging on the same answer. Most leaders are still pretending to debate it.

In Issue 51 of The Sunday Signal, David Richards MBE traces the line from Peter Drucker's most uncomfortable question to the modern Luddites, asks what the betting markets already know that boardrooms will not say out loud, and closes with the story of Leonie Hughes — a girl expelled from school at 15, now standing in a barrister's robes — and what AI cannot do.

This episode covers:

• Drucker's question and why every CEO is now avoiding it
• Microsoft's Rule of 70 and what the buyout really signals
• Polymarket's 90 per cent verdict on 2026
• The attack on Altman, and four lessons history teaches every boardroom
• Leonie Hughes and the people the new economy actually needs

This is not a transition. It is a replacement.

Read the full edition at thesundaysignal.ai. Subscribe free for the weekly Sunday signal on AI, work, power, and the future of Britain.

In January 2000, Pets.com paid $2.2 million for thirty seconds of Super Bowl airtime and sent a sock puppet. Nine months later the company was bankrupt. The sock puppet survived. The lesson did not.

Issue 50 of The Sunday Signal. This week I go back to the beginning — not the beginning of this newsletter, but the beginning of the pattern. Forty years of technology disruption, told without the survivor bias that distorts every boardroom presentation. Who lost in the mainframe era and why. Who lost in the web era and why. Who is losing right now in the AI era, and why the mechanism is identical every single time.

Then the investor's guide. Because the most dangerous sentence being said on earnings calls across the world right now is "AI is helping our business." Most of the people saying it are wrong. Some of them know it.

I give you a hot list — companies already dying, companies that cannot be touched, and the one question that separates them. Plus the Week 17 layoff tracker: Snap's CEO says the quiet part out loud, Oracle executes the largest single-company workforce reduction of 2026, and why an HR software company cutting its own staff to become "AI-first" is the most ironic signal of the year.

The sock puppet of 2026 is standing at the front of the room. The question is whether you act before the bankruptcy filing does it for you.

The Sunday Signal is a weekly newsletter and podcast on AI disruption, the future of work, and Britain's economic future. Published every Sunday at newsletter.djr.ai