Innovation is the opium of the boardroom

I needed some help with this one and turned to Google and Claude for pointers on Kurzweil's Law of Accelerating Returns, Marc Andreessen, William Nordhaus and the paper by Bloom, Jones, Van Reenen and Webb. None of whom I had heard of until I started asking the questions.

In 1983, my English A-level teacher asked me to write an essay about religion being the opium of the people. I wish I still had it. I cannot imagine how I managed to do the research to find out what on earth it was all about. I did write it; I wish I could remember how. If I get the time, I may try a visit to a library and try to do this the old fashioned way one day.

Although the essay is long forgotten, the phrase is remembered. When Marx called religion the opium of the people, he was not sneering. In 1843, opium was the only effective painkiller available, and the famous line sits inside a passage that calls religion the sigh of the oppressed creature, the heart of a heartless world. Religion, on his account, was both the expression of real suffering and a protest against it; it made an intolerable condition bearable, and in making it bearable, it removed the pressure to change it. The consolation was genuine, and that was the trouble.

Free market capitalism has its own opium, and it is not consumption or entertainment. It is innovation: the cure that is always coming. Inequality, climate, monopoly, stagnant wages, an ageing population, a crumbling grid; whatever the ailment, innovation is invoked as the remedy that makes structural reform unnecessary. You do not need to fix the system because it is about to fix itself. This is not a claim about any particular technology; it is a standing promise, and its function is to remove the drive to act now.

We do not have to reconstruct this faith from scattered remarks, because in 2023 Marc Andreessen wrote it down as a creed. The Techno-Optimist Manifesto is not a policy argument; it reads more like a sermon. It announces that its author has come to bring good news, repeats the words "we believe" well over a hundred times, closes with a list of patron saints, and invites the reader into the warm water. At its centre is the panacea claim stated as doctrine: no material problem, whether created by nature or by our own technology, lies beyond the reach of more technology. Starvation, darkness, cold, heat, isolation, pandemics, poverty; each is listed with the technology that answered it, and the sequence is offered as proof that the next answer is on its way. If you write like a preacher, you can hardly object when someone treats it as a religion.

The best response is not to laugh, because this faith is harder to shake than the one Marx described. Religion promised; innovation intermittently delivers. Antibiotics, cheap solar, the container ship, the Green Revolution that the manifesto rightly celebrates. A belief that occasionally pays out is the most durable belief there is, because every payout is read as confirmation of the whole doctrine.

William Nordhaus showed that innovators capture only about 2 per cent of the value their innovations create; the rest accrues to society. That may be true, and it is the strongest card in the techno-optimist hand, although the 2% figure is an average and should not be made into a slogan. But a fifty to one social surplus on the innovations that arrive tells you nothing about when the next one will arrive, or what you dismantled while you were waiting.

The problem is not that innovation is false; it is that it is deployed as a universal solution for questions that are more likely about power, incentives, and who bears the cost of the wait.

How the faith works

The faith rests on three things, each less visible than the last: a bet pretending to be a law of nature, a substitution, and a narrowing.

The first is a timing bet pretending to be a law of nature.

The manifesto leans on Kurzweil's Law of Accelerating Returns: advances become the tools that produce the next advances, so progress compounds and the spiral never ends.

The empirical view says something more awkward. Bloom, Jones, Van Reenen and Webb, in a paper bluntly titled "Are Ideas Getting Harder to Find?", showed that across semiconductors, agriculture and medicine, the number of researchers needed to sustain a constant rate of progress has grown enormously; keeping Moore's Law alive now takes roughly eighteen times the research effort it took in the early 1970s. The engine still runs, but the fuel consumption keeps rising.

Growth has been sustained not by accelerating productivity in the discovery of ideas but by throwing exponentially more people and money at the frontier. So, the real wager is never "innovation will deliver". It is "innovation will deliver on schedule, before this particular constraint binds".

That is a timing bet, and if the timing fails, the only thing standing between you and the consequences is whatever you kept in reserve. Organisations are poor at holding reserves even when they try, which deserves a paper of its own.

The second is a substitution.

In a leader’s mind, innovation and robust buffers are alternatives, with opposite cost profiles. Buffers are expensive now and pay off in crises you hope never come; they look like waste right up until they are the only thing standing. Innovation is cheap to invoke now and becomes someone else's problem if it arrives late. Give any executive short tenure and quarterly reporting, and the choice is obvious.

This is not stupidity or even unusual optimism; it is what happens when the costs are due now, and the benefits are due later, on someone else's watch. An organisation that trades its slack, redundancy and institutional memory for the promise that something clever will turn up has not reduced its risk; it has concentrated its risk into a single point of faith.

The third is a narrowing.

What is worshipped is not innovation in general but a specific kind: technological, product-shaped, purchasable, and originating from outside. The CIO who assumes someone, somewhere, is building the thing that will get the organisation out of trouble is not running an innovation strategy; he is running a faith in the vendor pipeline.

It is not that one kind of innovation is big and the other small. An endoclip, the device that can close a bleed in the stomach from the inside of an endoscope, is tiny: a titanium clip on the end of a scope. What separates it from the CIO's purchase is the direction of purpose. The endoclip spent twenty-five years being pulled toward a problem: a bleed that needed closing, refining, trialling, and certifying until it could be trusted inside a patient. The purchase is pushed from a solution: the vendor has built something, the analyst quadrant has blessed it, and the organisation goes looking for a problem worthy of the invoice.

The timing of the reward should be noted. The purchase produces its value at announcement: the slide, the vendor logo, the conference talk, the line on the CV. That value is real, and it is banked before adoption succeeds or fails; the sponsor has already been paid in reputation and has often moved on. The endoclip had to work to matter, but the platform only had to launch. Where accountability arrives at adoption, you get endoclips; where the reward arrives at announcement, you get shelfware and debt.

Nordhaus's ratio returns here, but it is inverted. Real innovation gives its creator about 2 per cent of the value and society the rest. Announcement innovation reverses it: the sponsor captures nearly all the value, in kudos, and the organisation keeps the residue and the decommissioning bill.

If this feels very anti-CIO, I guess it is.

A colleague writing about regenerative architecture wants to apply the same discipline, stop adding things, start taking things out, and create only what it can sustain.

Back to the CIO’s search for innovation, even the in-house version carries the same distortion as procurement because an R&D budget searches the space that an R&D budget can see. Meanwhile, the highest-return moves may sit in pricing, in the operating model, in removing an approval step that costs six weeks per decision, in killing a product line nobody will kill.

The organisation searches where its budget can see, not where the answer is.

What history shows

The techno-optimist has a ready answer to all this: human history. Two hundred years ago, it took most of the population to grow the wheat; now it takes almost nobody. Surely that is the machine working exactly as advertised.

Look at what did the work, though. The enormous gains in agricultural labour productivity were not the result of a single innovation. It was mechanisation, yes; it was also crop rotation, enclosure, futures markets, agricultural colleges, rural credit, seed certification. The invention of the tractor is what we remember, but the scaffolding of institutional invention is what made the tractor adoptable at scale. Some of the most consequential innovations in history were institutional: the double-entry ledger, the joint-stock company, the clinical trial, the weekend. Robust institutions are innovations that worked so well we stopped noticing they were innovations.

Therefore, the real divide is not between innovating and not innovating, but it is between additive and subtractive innovation, and here the manifesto makes a confession. Andreessen writes that the natural human drive "to make things, to gain territory, to explore the unknown" should be channelled into building technology and urges his readers to claim the technological frontier.

He is more candid than his critics usually manage to be. Innovation creates territories. The new system, the new capability, the new layer: each has a natural sponsor, because someone gets to own it. Subtractive innovation, the kind that removes handoffs and sign-offs and structure, has natural enemies, because someone loses the old thing.

The preoccupation with acquirable, additive innovation is not an intellectual error awaiting correction; it persists because it suits everyone who could change it. Leaders do not believe in innovation too much; they believe in the kind that spares them the kind that would cost them power.

Look at what the manifesto names as its enemies: risk management, trust and safety, sustainability, the Precautionary Principle. Every one of them is protective machinery, the checks and reserves that absorb shocks, and he denounces the lot as, by his count, sixty years of demoralisation, by which he means a campaign to turn people against technology. The substitution this piece describes is the stated programme: dismantle the protections and trust the innovation spiral.

What to do instead

A polemic earns its keep by saying what should replace the thing it attacks, so here is the programme, and none of it requires pessimism about technology.

• Audit the search space. Map where your innovation spend actually looks, and ask what fraction of it could ever find an answer that lives in the operating model, the pricing, the decision rights. If the answer is none, you are only ever going to find technology answers because that is all you are looking for.
• Give subtraction a sponsor. Removal needs its own budget line and its own career credit, because it will never win a political contest against addition on level terms. Somebody senior must be rewarded for what the organisation stopped doing.
• Count buffers as assets. Slack, redundancy and institutional memory are not waste awaiting optimisation; they are the material you will be innovating with when the wobble comes. An organisation that has traded all of them for promises has already chosen the crisis it cannot survive.
• Make the timing bet explicit. Every risk you defer to future innovation should carry a written arrival date and a named owner. If the cure is always coming, say when. A faith that must publish its schedule quickly rediscovers the value of insurance.
• Measure the function by obstacles removed, not things launched. Meaningful change comes at least as often from taking friction out as from adding capability in, and what gets measured is what gets sponsored.

None of this is an argument against technology. It is an argument against using the promise of it to avoid doing anything now. Marx's complaint was never that the opium failed to dull the pain; it was that the dulling postponed the cure. The question was never whether innovation would save us. The question is what we dismantled while we stood at the window, watching for it to arrive.

Afterword

I am uncomfortable with this article. I believe it makes some correct observations and conclusions, but all of human progress is built on ideas, and ideas become innovations when someone adopts them, and adoption depends on prepared ground: the means to reach people, afford it, trust it, use it, and fit it into how things already work.

So, we can understand how we have learned to trust innovations, and the groundwork is part of that.

And as I leave the Norfolk and Norwich University Hospital after visiting a sick mother, I think of the procedure she has had today, using those same endoclips. These were originally invented in 1975 in Japan, refined in Europe in the early 1990s, and globally commercialised just before the millennium. They continue to be developed, and no doubt one day will be replaced with another idea.

But look, twenty-five years from first use to the ward in Norwich. The essay's point stands even here, in the example that made me doubt it. Nobody plans around an idea that lands a quarter of a century out; no board would take that bet, and no patient can wait for it. What filled those twenty-five years, and what fills them still, is the other thing: the hospital, the training, the trials, the machinery that holds the line until the cure arrives. We have a lot to thank innovation for. We just cannot schedule it.