Until the 1960s, women couldn’t compete in Olympic events any longer than a sprint – and commentators declared that a marathon would kill them, or leave them unable to have children. Rubbish, of course. But when Kathrine Switzer signed up for the 1967 Boston Marathon, it wasn’t the distance that bothered her – it was the enraged race officials trying to assault her.   

Thanks to pioneers like Kathrine, women have made huge strides in long distance running – and are now challenging the times of men in the very races they were banned from for so long.  

This episode was first published in July 2024. I’m running the London Marathon next week – 26 April – to raise money for the Teenage Cancer Trust, and am grateful to everyone who has supported this very worthy cause.

[Apple] [Spotify] [Stitcher]

Further reading

On Bobbi Gibb

Ailsa Ross, “The Woman Who Crashed the Boston Marathon” JSTOR Daily 18 March 2013

Olivier Guiberteau, “Bobbi Gibb: The Boston Marathon pioneer who raced a lie” BBC Sport 29 August 2023

Brigit Katz “The incredible story of Bobbi Gibb, the first woman to run the Boston Marathon” Women in the World, New York Times, 20 April 2015

On Kathrine Switzer

Kathrine Switzer “The Girl Who Started It All” Runners World – excerpt from Kathrine Switzer Marathon Woman;

The Spirit of the Marathon by Gail Waesche Kislevitz Breakaway Books, 2002

“I ran with the men and changed history” BBC Outlook

On The Spine Race

Nick Van Mead “Montane Spine Race: 268 miles of pain”

Dave Lee “Spine Race 2013”

“Spine” (Amazon Prime documentary)

Jasmin Paris “Spine Race”

Episode 6 – Spine Race 2019

BBC Sport

BBC Scotland

Other sources

Roger Robinson “Eleven Wretched Women” – What really happened in the first Olympic women’s 800m. Runner’s World 14 May 2012 

Colleen English “Not a Very Edifying Spectacle”: The Controversial Women’s 800-Meter Race in the 1928 Olympics 08 October 2015

A Marathon Legend Revisited

Natalie Angier “2 Experts Say Women Who Run May Overtake Men” The New York Times 7 January 1992

Run Repeat State of Ultra-running

BBC More or Less

“We settled on a goal of 1,000 different photographs of every single object category,” she writes in her autobiography The Worlds I See. “One thousand different photographs of violins. One thousand different photographs of German shepherds.”

The database, ImageNet, was released in 2009, and Li started a competition for researchers to build the best image-recognition algorithms. A few years later, a graduate student named Alex Krizhevsky, advised by AI pioneer Geoffrey Hinton, trained a neural network on ImageNet — and blew the competition away.

Neural networks had been languishing for decades: a clever idea, but the computers had been too slow and the datasets were too small. But Li’s dataset was different. It had seemed foolishly, grandiosely, uselessly large; it turned out to be the perfect input for a neural net. This was an indication of the power of data combined with the power of neural nets. It was also vindication for Li’s idea of using human judgment to apply millions of labels to a vast collection of images. The lesson: if you can measure it, you can automate it.

But image recognition neural nets proved brittle in unexpected ways. A 2015 paper, “Deep Neural Networks are Easily Fooled”, asked a state-of-the-art system to classify example after example of pure static. “Robin,” said the network, with more than 99.5 per cent certainty, as it looked at random noise. “Armadillo.” “Peacock.” The problem was that the network had only ever seen meaningful images, and confidently identified meaning where there was none.

This is an example of the “jagged frontier” of AI capability, a term referring to the fact that AI models can be stunningly good at one task and then gravely disappointing at another, as with neural nets confronted with static.

That jagged capability is not a problem in itself. “All technologies are good at some things and bad at others,” says Joshua Gans, economist and co-author of Prediction Machines. It’s best to use can openers to open soup cans and hammers to drive nails into walls, and not the other way round. But, adds Gans, “the difficulty is that with AI, we don’t know which is which”.

This raises the question: how do we know that the AI is doing a good job? It was easy to see the problem when a neural net was labelling static an armadillo. But how impressive is the response to that request to create an image of Joan of Arc in the style of Edward Hopper? Did that agent actually make a restaurant reservation, or did it reserve nothing except a space in my calendar? Are the business plan and pitch deck I requested persuasive, or full of holes, or — perhaps the worst case — persuasive and full of holes?

The most problematic cases are the ones where it is hard to know whether the AI has done a good job, and expensive if it turns out that it has not. If AI writes buggy code or clumsy prose, that can be spotted and fixed. If the code contains hidden security vulnerabilities, the prose is packed with fabricated facts or plagiarised phrases, or the structural engineering calculations seem fine but the building will collapse in the first storm, that is a problem. It is still a problem even if the mistakes are rare and the average quality excellent. These difficulties only become more acute as AI becomes more capable, because more challenging tasks are often more difficult to evaluate.

Two new working papers address the tricky issue of verifying quality. In “Some Simple Economics of AGI”, Christian Catalini, Xiang Hui and Jane Wu (assisted, sometimes gratingly, by generative AI) propose the inevitable 2×2 matrix in which economic activity can be easy to automate, easy to verify, both, or neither. Automatable, verifiable output is the stuff that computers do for us. The non-automatable stuff remains reassuringly artisanal.

The difficult quadrant is where tasks seem easy to complete but are hard to check. Catalini, Hui and Wu call this the “runaway risk zone”. It is not a reassuring label and it is not meant to be. The problem of verifying quality is not a new one: think about building contractors, second-hand cars or a restaurant in a tourist hotspot. In such contexts, low quality often takes over the market like knotweed, because the best providers struggle to prove that they are the best.

Solutions include reviews, word of mouth, or long-trusted brands. (Not for nothing do familiar brands such as Durex and Trojan dominate the market for condoms. Nobody wants an unpredictable condom.) In big projects with high stakes, it can help to have the option to sue some counterparty with deep pockets. But none of these solutions is ideal, and the danger is that AI produces such vast vats of plausible slop that they outpace our capacity to check. Create enough hallucinated legal arguments, flawed engineering calculations and backdoor-ridden code, and the slop vats fill faster than our capacity to tell good work from bad.

In the second paper, “A Model of Artificial Jagged Intelligence”, Joshua Gans offers an analogy in which asking AI to perform a task is like trying to cross a river over a network of planks supported by occasional pylons. The jagged frontier is represented by the fact that some planks are long and wobbly, while others are short and sturdy. Problem one: even if the planks are typically sturdy, the wobbly planks will require most of your time and attention. Problem two: if you can’t predict in advance which planks will let you down, you may quite sensibly prefer to eschew the AI entirely and row yourself across the old-fashioned way.

As Gans rightly points out, Silicon Valley’s AI firms have mostly been trying to raise the average performance of AI systems — that is, to make all the planks sturdier. It might be better, instead, to focus on stiffening the wobbliest ones. But that assumes you know which they are, which points to a third approach: improve the predictability of the system. If you know in advance where the wobbly planks are, they’re not nearly as dangerous.

If.

Written for and first published in the Financial Times on 18 March 2026.

I’m running the London Marathon in April in support of a very good cause. If you felt able to contribute something, I’d be extremely grateful.

For ad-free listening, monthly behind-the-scenes conversations, our newsletter, and monthly bonus episodes including this one, please consider joining the Cautionary Club.

[Apple] [Spotify] [Stitcher]

Further reading

This episode relied on the books The Lysenko Affair by David Joravsky, The Rise and Fall of T D Lysenko by Zhores Medvedev, Lysenko’s Ghost: Epigenetics and Russia by Loren Graham, Stalin and the Scientists: A History of Triumph and Tragedy by Simon Ings, and Our Daily Bread: The Essential Norman Borlaug by Norman Vietmeyer.

Cancer is brutal for anyone. For teenagers and young adults, it can be isolating in a particular way: while friends are setting off on the adventures of early adulthood, they’re navigating a life-changing – and sometimes fatal – illness. TCT’s specialist nurses and teams work within and alongside the NHS to make sure young people feel safe, seen and supported. They advocate for them in systems that don’t always know what to do with a 19-year-old having chemo, bringing expertise and sensitivity when it’s needed most.

I’m running in memory of Winnie – a cousin and a childhood friend of my daughter. Winnie died of cancer on 22 April 2025. She was 20. Through countless challenges she was upbeat and determined. She kept studying for a law degree throughout her ten months of treatment, despite her university’s initial reluctance to support the idea. Her degree will be awarded posthumously – a testament to Winnie’s courage.

Throughout her treatment, Winnie and her family were able to rely on TCT clinical nurse specialists for support, advice and advocacy. Her family are enormously grateful to TCT, and when I asked which charity I should run for, they didn’t hesitate.

I’m grateful for the chance to try to complete the London Marathon, and I don’t plan to waste it. If you’re able, please donate to the Teenage Cancer Trust. Your support will help TCT’s nurses be there for more young people like Winnie.

—

If you’re curious about how my training is actually going… thanks! It’s been… fine. I’ve learned a lot. The toughest lesson, although perhaps not surprising, is that while having a marathon hanging over you can be a great motivator, it can also be a source of misery. It turns out I don’t really need a marathon to motivate me – but I might have had a bit more fun without the little voice in my head telling me I can’t do it…

I wrote about fitness trackers and algorithmic training programs here – all their amazing strengths and dangerous seductions.

And if you want to see me doing stupid stuff on a treadmill, on Facebook – it is the place for all displays of Unc Energy after all – the link is here.

In 1860, police officer Robert O’Hara Burke plans an expedition to map the ‘ghastly blank’ in the centre of Australia. Joining him is scientist William Wills, and a ragtag team of hires. Burke falls out with virtually everyone around him, and demonstrates an uncanny ability to make catastropically bad choices – from the equipment he brings to the route he takes. But even as the mission unravels, one final, simple decision could still save him.

For ad-free listening, monthly bonus episodes, monthly behind-the-scenes conversations, our newsletter, and more, please consider joining the Cautionary Club.

[Apple] [Spotify] [Stitcher]

Further reading

This script relied mainly on two books: Cooper’s Creek by Alan Moorehead, and The Dig Tree by Sarah Murgatroyd. The more recent Burke and Wills by Peter FitzSimons is a richly detailed but lightly novelised account. Thomas Schelling discusses coordination games in his book The Strategy of Conflict.

Since Hardy was writing in 1940, there was a touch of the humblebrag about this claim. Chemist Fritz Haber had created chemical weapons for use in the first world war. Engineers had produced artillery, tanks and strategic bombers. Oppenheimer and the other physicists would soon create the atomic bomb. There was a comfort in Hardy’s protestations of uselessness — but perhaps a false comfort.

In the 1970s, some basic ideas in supposedly useless number theory were deployed by Ron Rivest, Adi Shamir and Leonard Adleman. They developed the RSA algorithm, which enables public key cryptography, without which there would be no ecommerce. Cryptography is hardly valueless to the military, either. One never knows when useless knowledge will be useful after all.

Hardy’s number theory was not alone in being accidentally useful. In a famous article published around the same time — “The Usefulness of Useless Knowledge” (1939) — the head of Princeton’s Institute for Advanced Study, Abraham Flexner, made the case for apparently useless research. Flexner started with the radio and the radio telegraph — remarkable inventions for which many people thanked Guglielmo Marconi, the Nobel Prize-winning engineer.

Flexner argued that the “real credit” should go to James Clerk Maxwell and Heinrich Hertz, who had done the fundamental research. “Neither Maxwell nor Hertz had any concern about the utility of their work,” wrote Flexner, adding that Marconi contributed “merely the last technical detail . . . now obsolete”.

Some more recent examples have been gathered by the American Association for the Advancement of Science for its Golden Goose awards. Ten years ago, the awards recognised the Honey Bee Algorithm, which began with biologists painting tiny numbers on the backs of chilled (and thus immobile) bees, and then tracking the individual bees to figure out how they contributed to the hive’s search for nectar. Why? Because they wanted to know.

A couple of engineers became intrigued, figuring that maybe the bees had evolved a smart mechanism which the engineers might use to . . . well, do something. Perhaps they could use it to smooth the flow of traffic or suchlike. The bees had indeed evolved a clever approach, but the engineers couldn’t work out how to use it.

Finally, a computer scientist (Oxford, IBM) got in touch with the engineers, speculating that he had a problem to which they might have a solution. He was right. The honey-bee foraging system was adapted to spread viral and ever-shifting internet traffic across many different servers.

The Golden Goose awards also recognised the microbiologists who poked around in the geysers of Yellowstone Park to understand how some bacteria managed to thrive at very high temperatures. The scientists discovered heat-resistant enzymes — polymerases — that could survive near boiling point. This, quite unexpectedly, paved the way for the polymerase chain reaction — a way of amplifying genetic information made all too famous by the PCR test of Covid-19 fame, but one which has many other applications.

The Golden Goose awards do not exist in a political vacuum: they are explicitly designed to showcase the unexpected benefits of federally funded research in the US, and were meant as a rebuke to the earlier Golden Fleece awards, in which US senator William Proxmire would mock what he considered wasteful government spending — often on strange-sounding scientific projects.

Proxmire was not wholly wrong: some government projects are a waste of money, and some academics produce research of little value. But the lack of value is generally not because the research is “useless” but because the research is sloppily or even fraudulently done. Superficially interesting claims congeal on the surface of a steaming vat of confusion.

Unfortunately, politicians are not well placed to venture an informed opinion on the value of scientific research. The fact that research sounds silly or strange is no guide to its value. My own hunch — and it is just a hunch — is that it’s the research that seems obviously useful that is most likely to be polluted by bad science. The merely odd, purely curiosity-driven research is less likely to be tainted. Incestuous as it might seem, the people best placed to hand out funding for basic scientific research are other scientists.

This is not to say that society should just write a blank cheque to researchers. There are plenty of useful ways to guide scientific research.

One possibility is the use of innovation prizes, where funders specify a goal, and research teams are rewarded for achieving it. Examples range from the longitude prizes of the 18th century to the advanced market commitments that have been used to subsidise vaccine doses in the 21st century. Darpa’s grand challenge of 2004 and 2005 helped jolt life into the field of autonomous vehicles for a few million dollars in prize money.

Another possibility is to explicitly favour long-shot research with a high chance of failure but a real prospect of creating a major breakthrough. The economists Pierre Azoulay, Joshua Graff Zivin and Gustavo Manso compared grants made by the Howard Hughes Medical Institute against the more cautious approach of the National Institutes of Health. They found that both organisations got what they were asking for: a higher success rate for the NIH, and a mix of failures and breakthroughs for the HHMI.

A healthy scientific ecosystem needs both. And perhaps most of all it needs the odd-sounding, curiosity-driven research that no venture capitalist would dream of funding. The Nobel Prize-winning physiologist Archibald Hill once gave a public lecture at which a grumpy member of the public challenged him to explain what possible practical value there might be in his research.

“To tell you the truth,” replied Hill, “we don’t do it because it is useful but because it’s amusing.” That’s the spirit.

Written for and first published in the Financial Times on 11 March 2026.

I’m running the London Marathon on 26 April in support of a very good cause. If you felt able to contribute something, I’d be extremely grateful.

As the Victorian era dawns, modernisation erodes the old ways of life and poverty rises. In the unrest, an unlikely hero emerges, capturing the imagination of the countryside’s working class. He claims to be the new Messiah, and promises a better future. Despite his unconventional appearance and strange claims, his message resonates with the people of Kent, many of whom are willing to follow him into bloody battle.

For this Cautionary Conversation, Ian Breckon – author of Mad Tom’s Rising: The Revolutionary Mystic Sir William Courtenay and the Last Battle Fought on English Soil (US) – joins Tim to discuss a  forgotten folk hero and the dangerous power of belief in desperate times.

For ad-free listening, monthly bonus episodes, monthly behind-the-scenes conversations, our newsletter, and more, please consider joining the Cautionary Club.

[Apple] [Spotify] [Stitcher]

Further reading

Academics at UC Berkeley’s Haas School of Business have been doing ethnographic research into how technology workers are using generative AI. Some will tell you that ethnographic business research is both the worst kind of business research and the worst kind of ethnography, but I admit to a soft spot for this stuff. What the researchers found was the opposite of Adams’ morose Vogon guard: the minutes are amazing but the hours are terrible.

“In micro moments of prompting, iterating and experimenting, people talked about momentum and a sense of expanded capability,” researcher Xingqi Maggie Ye explained. “But when they stepped back and reflected on their broader work experience, a different tone sometimes emerged. They described feeling busier, more stretched, or less able to fully disconnect.”

These tech workers felt that generative AI was making them dramatically more productive and capable — but they were also trying to do more, voluntarily working longer hours, and hurtling towards burnout.

Are these ethnographic observations a glimpse of the future for the rest of us?

No doubt we shall find out, but while we wait, both economic theory and the history of technology have some things to teach us.

Theory first. Consider a freelance programmer, paid by results, who used to work 10 hours a day and suddenly finds that they can achieve the same results in two. Common sense might suggest that the coder will start to enjoy the pleasures of a two-hour workday, but economic theory is more ambiguous: the “income effect” suggests that the worker should work fewer hours because they can achieve so much by working so little. The “substitution effect” says that workers should work longer hours, as each extra hour yields bountiful rewards.

Then there is the question of what the new equilibrium will be once everyone masters the technology. As an analogy, imagine that a few alchemists discover how to turn lead into gold, but their method is rapidly being copied. They should make and sell as much gold as possible, as fast as possible, before the collapse in the gold market. Coders armed with brilliant AI agents may be in the same position: code as much as you can while you can still charge money to do so, because code may soon become as cheap as dust.

There is also a corporate dynamic to consider. It may be that nine out of 10 in-house programmers are about to be sacked, leaving a handful in charge to manage the coding agents. If so, the imperative is clear: to keep your job, demonstrate that you can out-code everyone else in the building. Winner-take-all dynamics are not a recipe for long lunch breaks and long weekends.

That’s the theory, but history has a few lessons for us, too. Visual aids were once produced by graphic designers and used on special occasions only; the invention of PowerPoint meant that highly paid and skilled professionals started wasting time making their own slides, badly. Email is vastly quicker and cheaper than a letter, but that simply means a profusion of low-quality, low-value messages bleeding into the evenings and weekends. The library photocopier allowed a generation of students to copy academic articles at a speed their parents could hardly have imagined — but it did not make reading, thinking or learning any faster.

In each case there was an astonishing increase in a narrow measure of productivity, but the overall effect was to distract from the real task at hand, to create a bloated pile of busywork, and to intensify the sense of productivity debt, with the list of tasks people felt guilty about not doing getting longer, not shorter.

What the UC Berkeley ethnographers found is strangely familiar. “Workers increasingly stepped into responsibilities that previously belonged to others,” they wrote. That’s the bad slide problem all over again.

“Because AI made beginning a task so easy . . . workers slipped small amounts of work into moments that had previously been breaks.” Everybody who lived through the rise of smartphones will nod in recognition.

“More multitasking. AI introduced a new rhythm in which workers managed several active threads at once . . . This created cognitive load and a sense of always juggling.” Well, yes — how many browser tabs do you have open right now?

I don’t mean to suggest that AI is useless or trivial, but there is a long history of time-saving digital technologies that at best make us more productive yet overwhelmed — and at worst, just make us feel overwhelmed.

Digital tools don’t have to work this way. The Nobel laureate economist Claudia Goldin coined the phrase “greedy jobs” to describe roles such as those in corporate law or investment banking where disproportionate rewards are paid to those willing and able to work long hours and be on call whenever required. She contrasts these with well-paid positions in pharmacy, paediatrics, primary care and veterinary medicine, where the jobs and the IT systems that support them have been designed to allow highly qualified practitioners to work limited hours and then hand over to an equally qualified colleague.

It’s not impossible to imagine AI agents being used to facilitate this handover process, but the discourse at the moment is of brilliant, idiosyncratic human conductors overseeing a frenetic orchestra of AI agents. Handover protocols sound less fun but may be a lot more useful.

And what are the rest of us to do while we wait for the wizards of Silicon Valley to stoop to building such prosaic tools? Todd Brown, a performance consultant and managing partner at Next Action Associates, has long espoused keeping an “Agenda” list for colleagues and important clients — working through the list face-to-face rather than firing off emails whenever something pops up. Now he does the same for ChatGPT, “with ideas for prompts”.

It may sound like an odd practice, given that generative AI — unlike a colleague — is always available. But it makes sense. Just because you can turn to AI at a moment’s notice doesn’t mean you should. There is something to be said for planning ahead before interacting with the AI, and for blocking out time without it — leaving space for the human in the loop to stop, to reflect and to breathe.

Written for and first published in the Financial Times on 4 March 2026.

I’m running the London Marathon in April in support of a very good cause. If you felt able to contribute something, I’d be extremely grateful.

In November 1979, Flight 901 departs New Zealand on a sightseeing journey over Antarctica, heading directly towards a volcano. When the plane vanishes, investigators are left with a mystery: how could a seasoned pilot miss a 12,000-foot peak? As they try to piece together the incident, conflicting stories emerge, key evidence disappears, and a troubling picture takes shape – one defined by human error, deceptive illusions, and the hunt for someone to blame.

For ad-free listening, monthly bonus episodes, monthly behind-the-scenes conversations, our newsletter, and more, please consider joining the Cautionary Club.

[Apple] [Spotify] [Stitcher]

Further reading

Many of the key documents are collated at the website https://www.erebus.co.nz, including the initial report of the Chief Inspector of Air Accidents and the findings of Peter Mahon’s Royal Commission. Mahon later wrote a book called Verdict on Erebus. Gordon Vette also wrote a book about the accident and his research into whiteout, called Impact Erebus.

In 2011, New Zealand journalist Paul Holmes worked with Jim Collins’ widow and daughters to write the book Daughters of Erebus. In 2019, Stuff and RNZ interviewed many of the surviving key figures for the podcast series White Silence. For a flavour of the debate that Erebus can still provoke, see forum threads such as this and this. James Reason comments on the Mahon report in his book Beyond Aviation Human Factors, co-authored with Daniel E. Maurino, Neil Johnston and Rob B Lee.

What happened is often known as the “conformity experiment”, but that is a misleading label for an oft-misunderstood study. Asch ran many variations on his experiment, and the most surprising and powerful lesson is not about the power of conformity, but about the power of disagreement.

Asch’s basic approach was to show two cards to a group of about eight people. One card had a single line on it: the reference line. The other card displayed three lines of different length. The task was a straightforward multiple choice, picking the line that was the same length as the reference line. This wasn’t hard; when people were asked to do this task on their own, they almost never made a mistake.

However, Asch was not asking people to do this task in isolation, but as a member of a group. Participants would be asked, one by one, to tell the rest of the group their answer. This made space for the possibility that experimental subjects would be guided not by their own eyes, but by the opinions of others.

The groups were asked to do this 18 times, but Solomon Asch had a trick to play. Everyone in each group was a confederate working for Asch, except a single unsuspecting experimental subject. This poor dupe would be sitting near the end of the line. The confederates had instructions to get the first two questions right and then unanimously agree on the wrong answer for most of the rest.

Imagine the jolt of surprise and anxiety as the experimental subject saw one person after another contradict the evidence of his own eyes. People felt real pressure to conform, with more than one-third of the answers matching the group’s delusion rather than the obvious truth.

Why? When debriefed, some people said they had changed their minds, figuring the group must be right. Others said they didn’t change their minds, but did change their answers, not wanting to spoil the experiment. Still others were staunchly independent, saying that they presumed the group was right and they were wrong, but felt a duty to call them as they saw them.

What fascinates me about Asch’s experiment is what happened when one of the confederates had been instructed to disagree with the group and give the correct answer instead. The answer: the spell of conformity was broken. People made only a quarter as many errors, with the error rate falling below 10 per cent. The pressure from the group had lost much of its power.

Even more brilliant was another variation in which Asch again instructed a confederate to disagree with the group. This time, however, the confederate was an “extremist dissenter”, giving an answer that was even more wrong than the majority consensus. The result? The experimental subjects generally gave the correct answer; their error rate was still below 10 per cent.

Asch had demonstrated three things. First, people will go against the evidence of their own eyes if contradicted by a unanimous group. Second, group pressure is much weaker if even a single person dares to disagree with the group. Third, and most remarkable: it does not matter if the dissenter is mistaken; dissent punctures group pressure either way. People are liberated to say what they believe, not because the dissenter speaks the truth but because the dissenter demonstrates that disagreement is possible.

I thought of Solomon Asch when I heard about a cookbook by Julia Child and Jacques Pepin, Julia and Jacques Cooking at Home. It’s full of the classics, but there are two very different recipes for each dish — one by Julia and one by Jacques. In the margins, each offers a jovial explanation of what the other cook has done wrong, why they made different decisions and what effect those decisions have on the final meal. It is, writes philosopher C Thi Nguyen, “the record of an argument — a rowdy conversation between friends”.

This matters because, as with Solomon Asch’s duplicitous experiment, it shows us that disagreement is possible. The two cases seem very different, not least because while there is only one correct answer to Asch’s visual perception test, there is more than one way to sauté a fish. Yet the disagreement is valuable either way, because it gives us permission to think for ourselves.

Many years ago I was involved in scenario planning for the oil company Shell. It was always a fascinating exercise, but I now realise that one of the most important strengths of the process was rarely discussed: there were always at least two scenarios, and all the scenarios were given equal status. This was Cooking at Home meets corporate strategy: the fundamental assumption was that there was more than one plausible future, and a rowdy conversation about the different possibilities unlocked a treasure chest of fresh thinking.

Charlan Nemeth is a psychologist and the author of No! The Power of Disagreement in a World that Wants to Get Along. She cautions against “contrived” dissent — for example, the Catholic tradition of having a “devil’s advocate” to argue against the canonisation of a putative saint. This sort of thing sounds good in principle, she argues, but in practice there is a limited benefit in a rote play-acting of disagreement. For one thing, everyone knows the devil’s advocate is just pretending, so nobody feels much pressure to persuade them to change their mind. “Role-playing,” writes Nemeth, “does not have the stimulating effects of authentic dissent.”

Yet some contrivances are better than others. Nemeth writes approvingly of an investment firm only making decisions after considering serious arguments both for and against a position. What makes this different from playing devil’s advocate? Perhaps the sense that the contrary arguments are not a game, but made in all seriousness.

Another contrivance is the idea of “red teaming” an idea — giving a group the task of trying to rip a new idea apart before that idea is adopted. Is this an empty ritual, or a serious practice? Depending on people’s intent, it could be either.

Contrived dissent is better than nothing, especially if the contrivance itself is taken seriously. But the most valuable form of dissent is authentic, even stubborn and brave. There is no substitute for finding one of those people who feel a duty to call things as they see them.

Written for and first published in the Financial Times on 25 Feb 2026.

I’m running the London Marathon in April in support of a very good cause. If you felt able to contribute something, I’d be extremely grateful.