Undercover Economist

The antisocial network?

Two economists have been collecting data to assess whether online friends are good for the soul. The quick answer: not really

Eliza was the first software to simulate human conversation. It was developed in the mid-1960s by Joseph Weizenbaum, a computer scientist at MIT. The most famous version mimicked a psychotherapist, thus hiding its conversational incompetence: “Men are all alike.” “In what way?” “They’re always bugging us about something specific or other.” “Can you think of a specific example?” “Well, my boyfriend made me come here.” “Your boyfriend made you come here?” And so on.

It was a clever program but Professor Weizenbaum, who died in 2008, was disturbed by the fact that several people seemed to find talking to Eliza genuinely therapeutic. Computers, concluded Weizenbaum, might not be terribly good for our emotional wellbeing.

These days, concern has moved to the amount of time we spend checking social networks – even if our online friends are, presumably, human beings. Two economists, John Helliwell and Haifang Huang, have been collecting the data necessary to assess whether such online friends are good for the soul. The quick answer: not really.

Helliwell and Huang analyse a Canadian social survey of 5,000 people called the “Happiness Monitor”, which measures how happy, stressed or satisfied with their lives people are, using a variety of standard questions. The Happiness Monitor also asks people how many friends they have in their real-life social network, as well as how large their online network is.

They find that having troops of friends is correlated with a sense of wellbeing. (As is common with such exercises, the direction of causation is unclear: perhaps happy people attract friends.) The effect is substantial: having twice as many friends is associated with the same increase in happiness as having a 50 per cent increase in income. But move the social network online, and larger networks do nothing for our happiness. Millions of digital sceptics will be unsurprised.

I am sceptical about the value of Facebook myself, but the most natural reading of Helliwell and Huang’s results is that a Facebook “friend” is not necessarily a friend at all, just a setting that tells software whose status updates to show us. The Happiness Monitor doesn’t even use the word “friend” when asking about the size of online social networks. (I have 73 Facebook friends and 65,000 Twitter followers, and it is not clear which represents the size of my online social network.)

Another study, by Fenne Deters and Matthias Mehl, published late last year in Social Psychological & Personality Science, asks a different question about our online socialising: how do we feel when we post status updates to Facebook? And how do we feel if nobody responds? Deters and Mehl ran a randomised trial with 102 students at the University of Arizona. The control group was given no specific instructions; the treatment group was asked to post more status updates “than they usually post per week”. Some ignored the instruction – but those who did not said they felt less lonely. It would be easy to over-interpret these results: the sample is small and there is something artificial about posting updates to Facebook in response to the request of an experimental psychologist.

The study is intriguing. It did not seem to matter whether anyone responded to the status updates. Perhaps people felt that they were being read even if there was no feedback; or perhaps responses came via email, text or face to face, unseen by the experimenters.

Or perhaps Facebook updates make us feel connected even though nobody out there is listening. That suggests a curious view of social networking: it may have little to do with true socialising. We may simply feel satisfied with the illusion that someone is paying attention. Joseph Weizenbaum, the creator of Eliza, would not have been surprised.

Also published at ft.com.

Patently a stitch-up

Are smartphone patents helping innovation – or strangling it?

I’ve been thinking about a remarkable industrial designer, a man who built on previous innovations to produce an indispensable product and a corporate titan. Not Steve Jobs, of course, but Isaac Singer, the developer of the first commercially successful sewing machine.

Singer and Jobs have something else in common: their products both became embroiled in bitter legal wrangling over patents. Jobs told his biographer, Walter Isaacson, that “I’m going to destroy Android, because it’s a stolen product. I’m willing to go thermonuclear war on this.” Singer threatened to kick Elias Howe down the stairs when Howe demanded cash from him for infringing Howe’s patent.

The courts eventually backed Howe, but many other patents existed and soon the entire US sewing machine industry seemed more interested in suing rivals than selling sewing machines. That will surprise nobody who pays attention to the legal battles over smartphone patents – and many observers have come to the conclusion that patents are not helping innovation, but strangling it.

Is that true? Patents give a temporary monopoly to inventors. This will raise prices and constrict supply, which is bad. But in the long run the patent should promote innovation by encouraging inventors to develop new ideas. This is a balancing act, not a matter of moral or practical absolutes.

But the sewing machine patent wars, and the modern smartphone litigation, show us that the balancing act is more complex when a technology may embody many different patents. According to one well-publicised estimate, there are 250,000 patents relevant to a modern smartphone. Even if the number is one-tenth of that, it suggests an impossible thicket of intellectual property through which a company must hack to bring a cool new product to market.

A key issue is something called the hold-up problem. If a $1bn product depends on 1,000 patents, it is clearly impossible to pay the typical patent holder more than $1m. But any patent-holder could try to extort many times that amount by threatening to block the whole project.

Large firms have responded to this problem by buying or developing large collections of patents. This gives them the ability to launch countersuits, and that threat should make rivals reasonable. But although defensive patenting looks like a pragmatic solution, it has costs and limits. The wave of defensive applications swamps patent offices, which means more poor-quality patents and longer delays.

“Patent trolls” – a derisive name for companies that make money purely from their patents – have less to lose in a patent war but although some are legitimate, others are extortionists. And while established players may reach cosy understandings, a young company with a new idea may find it impossible to break into a market that is thick with defensive patents. If only the big boys can play the patent game, innovation will suffer.

It’s clear that some industries are plagued by nuisance lawsuits. According to a survey by economists Bronwyn Hall and Dietmar Harhoff, firms complaining of infringement win more than half of court cases involving non-software patents, but only 13 per cent involving software patents. That suggests software suffers from weak patents and aggressive litigation, perhaps in the hope of extorting a settlement. It is not a good sign.

Yet we should not lose hope. The sewing machine companies eventually pooled their patents and competed on the high street rather than in the High Court. And despite apparently endless litigation, there seems to be plenty of innovation in the smartphone industry. The patent system is not a total failure – but on present evidence it is hardly a total success.

Also published at ft.com.

A hire truth

New research shows with horrible clarity what a wretched trap long-term unemployment is becoming

When my friend Nicola quit her job in educational publishing, it was for understandable reasons: one of her colleagues was a bully, her boss was providing little support and work had become miserable. (I’ve changed her name.) She had a strong CV, a husband with a steady job and young children to spend time with so she decided to resign before finding a new job. She’s now been unemployed for a year and a half.

So many economies have been depressed for so long that Nicola’s predicament is common. But new and unpublished research from a young Lebanese PhD student, Rand Ghayad of Northeastern University in Boston, shows with horrible clarity what a wretched trap long-term unemployment is becoming.

(April was a good month for economics students elsewhere in Massachusetts: Thomas Herndon of the University of Massachusetts, Amherst, delivered a reputational kneecapping to Harvard professors Carmen Reinhart and Ken Rogoff when he discovered a spreadsheet error in their much-trumpeted paper “Growth in a Time of Debt” as part of a course assignment.) (Here’s our coverage on “More or Less” – MP3.)

Ghayad used a computer program to generate job applications that were standardised but varied along a few parameters – whether the imaginary applicant had worked in the relevant industry, had hopped a lot between jobs, and how long the applicant had been unemployed, if at all. Ghayad mailed 4,800 of these CVs off to apply for 600 vacancies, chosen to reflect a variety of city locations, seniority and industry. He then recorded which applications were offered an interview.

Unsurprisingly, candidates with recent relevant experience were at an advantage, and a history of job-hopping did not help. But what was astonishing was the effect of long-term unemployment. Applicants with experience from the wrong industry who had been unemployed for 14 weeks or less were more than three times as likely to receive a call from the employer than applicants with experience in the right industry but who had been unemployed for six months or more. Regardless of sector, employers are, apparently, more interested in shunning the long-term unemployed than in looking for relevant experience.

The implications for Nicola and others like her are clear. But it is puzzling that employers put such emphasis on avoiding the long-term unemployed. A long period of unemployment is a negative signal, of course – it suggests that the applicant might not have been trying, might be getting rusty or might have been interviewed and rejected several times by other employers. But quite why it is such an overwhelming stain on a CV is not clear. Six months isn’t even a long time – in Europe, you don’t even count as long-term unemployed until you’ve been looking for work for a year.

Ghayad believes the evidence suggests that employers are using long-term unemployment as a quick – perhaps even automated – reason to disqualify applicants for oversubscribed vacancies. He thinks the policy implications are that the US needs more economic stimulus, because only a surge in employment would force employers to give the long-term unemployed a chance.

What’s certainly clear is that this reluctance to hire the long-term unemployed is one way in which the recession is leaving long-lasting scars. This has happened before: one of Margaret Thatcher’s less trumpeted achievements was a rise in the equilibrium unemployment rate that lasted into the mid-1990s. One likely cause was simply the number of people who became severed from the labour market in the deep recession of 1981.

I asked Ghayad, 27, why he’d chosen a PhD in economics. He had a simple answer: “I graduated during a recession and I couldn’t find a job.” Silver linings, for some.

Also published at ft.com.

An evidence revolution

Teachers are better placed than anybody to generate new research questions, based on years of observation of subtleties that would escape any educational statistician

A while ago, my daughter’s school explained to parents that they were reorganising literacy classes, moving to mixed-ability groups. My wife asked a simple question: “What is the evidence behind the decision to change?” The result was blank incomprehension. Evidence?

Only a couple of generations ago, most doctors had a similar attitude to evidence-based practice. Fancy statistical procedures were thought to be no match for experience, especially since every case was unique. Randomised trials were all very well in theory, but unethical in practice. That scepticism has now been completely transformed in medicine: evidence trumps seniority, while individual judgment is bolstered by online libraries full of careful analysis.

Teachers have allowed themselves to be left behind in the evidence revolution. I sympathise with the profession which is constantly second-guessed by parents and school inspectors. Teachers have grown used to fad after fad being hurled at them from the Department for Education.

But I agree with Ben Goldacre – epidemiologist and author of an excellent polemic on evidence-based education commissioned by the DfE itself – when he argues that if the teaching profession embraced the evidence-based approach, it would enhance rather than diminish its independence from government. The facts, after all, rarely slot neatly into political ideologies. At a recent dinner organised by the Wellcome Trust, the British educational establishment – with, alas, only one teacher present – discussed the issue. The longstanding tension between teachers and governments was clearly recognised as an obstacle – but it’s one that can be sidestepped if teachers themselves seize the evidence agenda.

This is about much more than simply running randomised trials comparing different approaches to teaching. Consider the situation in clinical practice, as outlined by Professor Jonathan Shepherd, who argues that the secret is a tight plait of research, practice and continuing education. Trainee doctors are taught by practising clinicians. Those clinicians are also researchers, whose research agenda is closely influenced by their clinical experience. Research networks link together qualified researchers with GPs who have patients and research ideas. And once qualified, doctors continue to have the latest evidence pushed under their noses in the likes of the British Medical Journal.

In short, evidence-based practice in medicine isn’t a case of doctors, brainwashed into believing whatever clinical trials tell them, passively awaiting instructions. It’s a two-way street, where some of the best ideas for research are suggested by practitioners, and best practice spreads sideways from clinician to clinician rather than being handed down by diktat. There is nothing fundamental about education that makes this impossible – witness the “journal clubs” in Singapore and Shanghai, where teachers discuss and evaluate the latest research.

One can see why Dr Goldacre calls this a “prize”. Teachers are better placed than anybody to generate new research questions, based on years of observation of subtleties that would escape any educational statistician. There is, at last, some institutional support: the Institute for Effective Education at the University of York, for instance; or the Education Endowment Foundation, two years old this month, which is already running 50 randomised trials in schools, with a grant of £125m from the DfE.

“Trust me, I’m a doctor” was never an excuse for not collecting evidence. And “trust me, I’m a teacher” is not an excuse today. But being a teacher is a superb vantage point for building an evidence-based education system. It is an opportunity that teachers need to seize.

Also published at ft.com.

A brief lesson in letter-writing

New research by the Financial Conduct Authority shows how simple, low-cost research can yield substantial gains

This column will tell you:

• How to write an email that will produce the results you want;

• Why the scientific method can produce practical insights, quickly.

How can I promise this? I’ve just been reading up on the latest research from Paul Adams and Stefan Hunt of the Financial Conduct Authority – the UK’s shiny new financial industry regulator.

I am delighted to report that they’ve been trying to figure out how to write more effective letters. Better yet, to this end they have conducted a large randomised trial. The wonderful conclusion: science can help us write more clearly.

The letters in question were being written by a company to 200,000 former customers. The letters raised the prospect of a mis-sold product and offered the possibility of a refund. The question was: which wording would prod customers into claiming their money? The FCA researchers, in co-operation with the company in question, designed variants of the letter, mailed each variant to 1,000 customers, and compared the response rates.

The original letter looks brief and to the point, but on closer inspection it buries the key detail: the customer may be entitled to a refund. Response rates to this letter were dismally low, less than 2 per cent. Perhaps that is no surprise – the recipients were ex-customers with every reason to throw the letter in the bin, unopened.

The experiment did the sort of thing that profit-seeking companies have been doing for many years: it tested seven different tweaks to the way the letter was written or presented. There are 128 different ways to combine these seven tweaks, and with each of 128 variants sent to 1,000 customers the experiment had a lot of power to pinpoint even quite small effects.

Here are three tweaks that made little difference: printing “important: please read and act quickly” on the envelope induced a minuscule extra response; adding the regulator’s logo achieved nothing; using the company CEO’s name and signature instead of “customer services team” actually dissuaded people from responding.

But four other tweaks had substantial effects: first, cutting a paragraph of waffle that had helped to bury the message about the refund; second, pointing out that a five-minute phone call would suffice to make a claim; third, sending a follow-up letter. And twice as large as any of these effects was adding a couple of bullet points in bold at the top with the key message: you may deserve a refund; if so, call us.

This research has been marketed by the FCA as “behavioural economics exploring how people make financial decisions”, but like similar work on collecting fines and taxes conducted by the Cabinet Office’s Behavioural Insight Team, it is simpler and more pragmatic than that. There is no behavioural theory at play here, and nor do we really gain any insight into why consumers are reacting the way they do. But that’s fine. Simple, pragmatic research is a sensible thing for a regulator to be doing – and the cost of this kind of experiment is tiny relative to the potential gains.

True, the results are unsurprising: say what you mean; be brief; ask for action; follow up if you hear nothing. But this is important. It’s important because it shows a regulator with an interest in learning and improving, and it’s important because while the advice is commonsense, it’s advice that many people – and even more bureaucracies – fail to heed.

And even if the findings seem obvious, the effect is huge. The best letter received seven times as many responses as the original one. The three best tweaks each made more of an impact than the decision to send a letter at all rather than nothing. Brevity and clarity matter: advice worth taking next time you write an email.

Also published at ft.com.

Microjobs have a massive potential

A new company aims to pair the least-skilled and most excluded workers with jobs that need doing on a more industrial scale

The internet has a wonderful habit of creating new markets. eBay is the most famous and enduring example of bringing together people with something unusual to sell and other people who want to buy it. (In 1995, its creator, Pierre Omidyar, listed a broken laser pointer for sale; when he received an offer, he repeated his warning that the thing was broken. The response: the buyer collected broken laser pointers. Bingo.)

More recently, sites like Airbnb and TaskRabbit allow people to rent small spaces, or work at small tasks. While they are commercial, they retain an emphasis on personal connection and community. Rising to prominence in 2012 was Kickstarter, which helps people with projects raise money in small chunks from enthusiastic backers.

A new company, Upstart, offers fundraising for people who don’t even have projects. The typical Upstart is a precocious Ivy League graduate, all set to walk into a job at Google, Goldman Sachs or Exxon, but instead seeking backers to provide them with the financial security to be an entrepreneur.

All this is exciting. But oddly, despite the prodigious ability of modern computers to process data, each of these new sites works on a very human level. Every TaskRabbit task, every Kickstarter project is unique; the website merely facilitates personal connections.

“It’s all a bit 1995,” says Wingham Rowan, a technology journalist turned entrepreneur. And while the slick, user-friendly new sites are a world away from the text-based blue hyperlinks of the early web, Rowan has a point. Underneath the glitz, TaskRabbit operates much like an efficient small-ad board at your corner shop.

Rowan is now strategic director of a company called Slivers-of-Time, and he sees a missing opportunity: to bring the sophisticated market-making technology to the least glamorous end of the British labour market – to the poorest, least-skilled and most excluded workers.

A typical Slivers-of-Time worker is looking for a couple of hours’ work here, a couple of hours of volunteering there, either because personal circumstances make it hard to commit to regular hours, or because a full-time job isn’t available. A Slivers employer could be anything from a supermarket to a local council, a call-centre to a café, looking for a little assistance with the lunchtime rush, a product launch, or staff holidays. Slivers creates a searchable database, almost in real time, to allow an employer to find suitable workers at short notice, for a reasonable rate and with the right qualifications – anything from a food safety certificate to a driving licence.

This approach is finding some takers – Tesco and several local councils use the technology. But it is slow going. Slivers-of-Time is less cute than the “hire someone with a nice photo to pick up some dog food” approach of TaskRabbit. It aims to pair unused labour with jobs that need doing on a more industrial scale. And as Rowan argues, it cannot reach its potential without a lot of government work to sort out the regulatory environment.

Benefits need to be reformed so that doing a few hours’ work a week is encouraged rather than penalised. (In principle, the new universal credit should allow this. In practice, it’s not easy.) Governments also need to reform their information infrastructure: providing secure and selective access to databases so that driving licences, criminal records or professional qualifications can be checked instantly.

Yet such markets are potentially very important. Development gurus have long emphasised unlocking the potential of the poor by cutting red tape, establishing property rights and providing access to affordable credit. But jobs are typically what lift people out of poverty. And if it’s possible to develop microcredit in Bangladesh, it should be possible to develop microjobs in Birmingham.

Also published at ft.com.

Rich pickings for scientists

Should we feel queasy about the growth of abstract, deeply technical thinking in finance?

Amidst all the glories of belle époque Paris, the troubles of an aspiring physicist can’t have amounted to much. In 1888, Louis Bachelier had graduated from school with excellent grades and high hopes of studying at the elite grandes écoles – until both his parents died suddenly, leaving Louis responsible for his sister and his three-year-old brother. He ran the family wine business for a while, was drafted into the army, and by the time he extricated himself from that and sold up, he was too old to do anything but study at the less-prestigious University of Paris. With his siblings to support, his study of physics had to be nocturnal. By day, he worked at the Paris Bourse.

Bachelier’s enthusiasm for physics quickly spilled over into finance, and he began to ponder the processes that governed the price of government bonds. He developed the idea of the random walk; he wrote up his ideas and defended a dissertation on the physics of financial speculation. He was almost laughed out of academia, and only secured a PhD at all with the help of a powerful mentor, Henri Poincaré.

Bachelier was ahead of his time even as a physicist. His research anticipated by five years a hugely important paper on Brownian motion by a young chap called Albert Einstein. But as a financial thinker, he was more than half a century ahead of the curve. His early research was discovered by the mathematician Leonard Savage and economist Paul Samuelson in 1955, a decade after Bachelier died. The first person to improve upon his ideas was arguably the physicist Maury Osborne, fully 59 years after their publication.

The sad story of Bachelier is told in an excellent new book, The Physics of Wall Street (US) by James Owen Weatherall. The role of physicists in finance is now a commonplace, even if financial physics is, like its founder, not quite academically respectable. But modern financial physicists – the “quants” – differ from Bachelier in one respect: some of them have made an awful lot of money. The most successful is Jim Simons. He was once a serious player in the high physics of string theory. Now, as founder of Renaissance Technologies and its flagship Medallion fund, he is one of the richest men on the planet.

Should we feel queasy about the growth of abstract, deeply technical thinking in finance? Weatherall is a cheerleader. For one thing, he says, Medallion returned more than 70 per cent in 2007 and 80 per cent in 2008. That proves quantitative finance can be privately profitable –but not that it is profitable in general, or socially desirable.

A more convincing defence of financial physics, and the sophisticated statistical analysis it deploys, is that it provides fresh perspective, revealing patterns that had been missed. A famous example is Didier Sornette’s use of stress analysis to predict not only earthquakes and failures of pressurised fuel tanks, but also severe crises in financial markets.

Still, physics alone will not be enough. Understanding the financial crisis, or the economy itself, should be a multidisciplinary effort. Unpicking the flaws behind stinky structured finance products required somebody to take an interest in messy real-world details. No purely academic approach would have delivered that.

Unfortunately multidisciplinary work remains difficult. Even the physicists have trouble being taken seriously by economists, despite the latter’s respect for mathematical sophistication. As for psychologists, sociologists, anthropologists … It’s a tough thing for any self-respecting economist to swallow advice from disciplines that appear to deploy such vague theories.

Still, nothing concentrates the mind like the potential to make money. And nothing fosters interdisciplinary work like a problem that scorns boundaries. Finance offers both the problem and the potential. Watch this space.

Also published at ft.com.

What Oxbridge can learn from YouTube

The British educational establishment should ignore online open courses at its peril

A couple of years ago, I showed my daughters a video put online by the Khan Academy, which has become famous as a pioneer in open-access education. The video was an amateurish but charming explanation of basic arithmetic. We had fun but the girls were not transformed into mathematical prodigies. Their mathematical education remains the sole responsibility of a rather traditional school in North Oxford. The only thing YouTube has taught them is how to draw manga cartoons.

That experience would not surprise the British educational establishment. Massive Online Open Courses (Moocs) are all the rage but the top universities seem to regard them as mere amusements, unlikely to threaten traditional methods, which may be costly but are exclusive and of excellent quality.

The vice-chancellor of Cambridge university, in a speech in January, said that online courses would “challenge the nature of higher education” but that they would not change what happened at Cambridge.

Educational expert Karan Khemka seems to agree, explaining in this newspaper’s comment page that the Mooc approach would eventually improve higher education, but “through incremental change rather than massive disruption”.

I am not so sure. Clayton Christensen of Harvard Business School has become famous for his explanation of how “massive disruption” tends to happen. A new technology appears, and it’s cheap and cheerful. Examples: small hydraulic shovels when the standard was large cable-driven mechanical diggers; small steel mills devoted to melting down scrap metal; slow, low-capacity hard drives in a world of mainframe computers. The dominant players in the industry look at the upstart technology and ignore it. This isn’t just a case of snobbery: their customers don’t want the new cheap technology either. They are already paying for a high-quality product that fits their needs perfectly.

The cheap technology is embraced by new entrants, who supply a totally new customer base. Construction companies didn’t want cheap, weak hydraulic shovels but landscape gardeners did. And then, of course, the new technology got better and better. Eventually it overtook what the incumbents had to offer and, despite all their advantages, they had absolutely no idea how to make or deploy the upstart technology that had become the state of the art.

My colleague Michael Skapinker has dismissed Clayton Christensen by providing a (perfectly reasonable) list of the ways in which a face-to-face learning experience beats a YouTube lecture hands down. It is most unlike Michael to miss the point.

Of course, a few online videos, chat rooms and multiple choice questions pose no threat at all – for now. Why go online when you can receive expert tuition in small groups, receive an exclusive stamp of achievement on your CV, and still have time to get drunk with the future masters of the universe? Online courses have little to offer current students at the world’s top universities.

That is why they are such a disruptive threat: if Oxford and Cambridge ignore Moocs now, what happens when the digital component of education has evolved to become indispensable?

If “content on demand” isn’t a killer application, then the ability to measure each student’s progress and tailor courses accordingly may well be.

The wise move has to be to follow MIT and Stanford, and indeed the UK’s Open University, embracing Moocs not for what they offer now but for what they might one day become. It is time for the UK’s greatest educational establishments to learn a few lessons themselves.

Also published at ft.com.

Geoengineering, a monster of our own making?

The core case against the science is a radical uncertainty about its consequences but it would likewise be irresponsible to turn our backs on it

While holidaying in the Alps with Lord Byron and Percy Bysshe Shelley, the young Mary Godwin was stranded indoors by unending rain. Stuck for activities to fill the days, she dreamt up a horrific story of a brilliant scientist whose hubris has tragic and unpredictable consequences. She later became Mary Shelley, and the story later became Frankenstein.

We now know that the dreadful weather of 1816 was induced by the vast eruption of Mount Tambora in Indonesia. Two centuries later, the use of imitation volcanic clouds is being seriously contemplated as an antidote to global warming. The word “geoengineering” is on the lips of the world’s atmospheric scientists.

The trigger for the discussion of geoengineering was a 2006 article by Paul Crutzen, a Nobel laureate and expert on the ozone hole. Many scientists share his concern that substantial climate change can no longer be prevented: we are emitting too many greenhouse gases and our plans to stop are tardy and timid.

Geoengineering proposals fall into two broad categories: we can try to remove carbon dioxide from the atmosphere, or we can try to reflect sunlight away to counteract the greenhouse effect that carbon dioxide produces.

Both ideas are superficially tempting. Carbon dioxide lingers for many decades, which means that, even if we stopped pollution tomorrow, warming would continue. Removing carbon dioxide would allow us to undo past harms more quickly and might prevent some tipping point being reached.

Solar radiation management – by creating reflective fluffy clouds to screen the dark oceans, or by using the volcano effect and pumping sulphur particles into the stratosphere – has its own attractions. For one thing, it seems absurdly cheap. So what’s the catch?

The catch is obvious enough. The world’s climate is complicated and we don’t really know what the consequences might be of interfering with it. We can guess at a few: some modelling suggests that a stratospheric sulphur shield could lower global temperatures to where we want them, but would not prevent the oceans acidifying, might affect the monsoon in India, and would cool the tropics while failing to cool the poles.

There are other risks, of course. What if geoengineering becomes a weapon? Clive Hamilton’s otherwise useful book, Earthmasters, is marred by dark mutterings about the connection between geoengineering and Lawrence Livermore National Laboratory, a centre for nuclear weapons research. This doesn’t worry me much. We have far simpler ways to obliterate our enemies.

No, the core case against geoengineering is a radical uncertainty about its consequences. But this cuts both ways. Global warming is a threat not only because of the likely scenario in which the climate changes in harmful ways but we adapt; but in the less likely (but plausible) scenario in which some runaway process makes the planet uninhabitable as we know it. For example: reflective ice melts, exposing dark oceans that absorb heat; warming accelerates; methane is released from the melting tundra; methane exacerbates the greenhouse effect; repeat.

While it would be irresponsible to rely on geoengineering to get us out of our present fix, it would also be irresponsible to turn our backs on the possibility that it might one day prevent catastrophe. Geoengineering experiments are, in any case, already happening – and they are cheap enough for a rogue nation or even a rogue Bond villain to have a go at something quite ambitious. It is time to start thinking about this, and quickly. As Mary Godwin realised, science plus overconfidence can produce an awful mess.

Also published at ft.com.

A hateful abuse of algorithms

Putting the blame on badly supervised computers for releasing offensive products is not much of an excuse

“Keep Calm and Hit Her”. At first glance the T-shirt design seemed to be yet another example of the way violent misogyny has become a joke for some – a literal punchline. No wonder that earlier this month the internet was up in arms at the discovery of such a T-shirt, not from a street trader or a Soho sex shop, but on Amazon.co.uk. And “Keep Calm and Hit Her” wasn’t the only example. There was also “Keep Calm and Rape A Lot”.

When we see such T-shirts for sale through the website of one of the UK’s largest retailers, we naturally imagine that somebody designed the shirt and some buyer at Amazon said, “We’ll take it.” The fact that similar jokes are a staple of the stand-up comedy circuit makes it all the easier to believe that.

But “rape a lot” is not only outrageous, it’s an odd phrase. And what about “Keep Calm and Skim Me”? This was also among the hundreds of thousands of T-shirts available from Solid Gold Bomb, the company which sold them via the Amazon website. (The T-shirts are no longer for sale.) Solid Gold Bomb issued a profuse apology and delivered an intriguing explanation: it wasn’t that somebody had decided these T-shirts would be good business – it was that the T-shirt slogans were being suggested by a computer run amok.

The T-shirts didn’t exist in physical form: a computer algorithm suggested a range of different slogans, and if any of them were purchased the T-shirt would then be printed to order. Now you might well ask what kind of computer algorithm would include the word “rape”; or why Solid Gold Bomb didn’t check the output of the computer (or did check, but didn’t pull the offensive logos). Well, indeed: “my algorithm made me do it” is not much of an excuse here.

The idea that men might crack jokes about rape is all too familiar. The idea that computers are algorithmically generating entire product lines is a new one on me. And while the algorithm excuse is unpersuasive in this case, the algorithms are coming. Search Amazon.co.uk for “Philip M. Parker”, a business school professor who has patented software to write dictionaries, benchmarking studies and other narrow genre books: I saw 113,000 results.

What next: algorithmic blasphemy? Algorithmic child pornography? Hard to say: by their nature, algorithmic processes tend to produce surprises. They can also be a force for good. Using the blind power of variation and selection, computers can design a better nozzle, a better surfboard, even a better campaign to persuade people to quit smoking. The key insight is that a very high error rate is acceptable because the successes are retained and the errors quickly discarded. But what if one of the errors goes viral instead?

One solution is to rapidly remove rogue results. But that introduces a different problem: when the environmentalist author Mark Lynas published The God Species, the book was briefly removed from the Amazon website during the launch campaign. Amazon’s site reported that it had received complaints that the product was “not as described”. Lynas suspected sabotage from his political opponents, and remains unsure as to what really happened. Complaints-based takedown procedures can themselves be abused in a world full of algorithms. (Disclosure: Lynas is a friend.)

Another option is more diligent pre-screening of results. That would have worked for the rape T-shirts. But it, too, has its own costs: millions of customers looking for legitimate niche goods may find that online retailers cannot be bothered to check the long tail of products, and turn back to the mass market.

As I watch my daughters grow up, I suspect they will have even more serious feminist concerns than offensive T-shirts designed by badly supervised computers. But something tells me we haven’t heard the last of algorithmic products.

Also published at ft.com.