Soon after the pandemic hit, the world’s airlines turned off their pricing algorithms and resumed pricing flights manually. Everything the software had learned from people’s past behaviour was suddenly rendered irrelevant.

The software had been created for a world of discretionary travel where demand was elastic. If a plane seemed likely to leave half-empty, the software dropped prices to fill remaining seats. In March this once-efficient approach failed spectacularly. The few people who were still flying were doing so only in desperation: everyone else was unwilling to travel at any price. Far from reducing prices to respond to a drop in demand, it now made sense to hike them.

It is sometimes said that insanity is doing the same thing again and again and expecting a different result. But doing the same thing under different circumstances while expecting similar results is no less a form of derangement. The first behaviour is a characteristic of alcoholics; the second is a hallmark of bureaucrats, ideologues, management consultants and devotees of process automation. It is most prevalent among the more educated, for whom theoretical neatness is a major status marker: the very people who design systems and software.

(I changed from a Remainer to a Leaver for these very reasons. I had voted Remain, but was so alarmed by the monolithic thinking of hardcore Remainers, for whom there seemed no problem to which further integration was not the answer, that I switched sides. The problem with the EU, it seemed to me, was the same as the problem with the Archduke Franz Ferdinand’s car: it’s very difficult to put it in reverse.)

Large institutions are prone to a kind of conceptual locked-in syndrome — and the larger the organisation, the worse the effect. The book Seeing Like a State, by James C. Scott, is a wonderful account of this problem. As evidence, consider the fact that we are proceeding unthinkingly with HS2 despite a once-in-a–generation change in travel habits — especially among business travellers.

Now, just to be clear, I think it possible that the widespread adoption of remote-working and video-conferencing might strengthen the case for HS2. (In time, the number of people who live and work elsewhere, but who travel to London once a fortnight or so, might rise.) But shouldn’t we re-run the numbers now? Perhaps HS2 is now a good idea, but not as good as rolling out broadband and 5G to remote areas, or allowing the Cornish to make 2G calls without standing on a hill?

As I have written before, the magical property of business and free markets is that it is the only sphere of human activity where you get paid to change your mind. One tech investor phrases his philosophy as follows: ‘All you need do is make a list of the assumptions in a business sector, and then find the ones which aren’t true — or which won’t be true in two years’ time. You’ve now got a business.’ It is this competing market for conflicting ideas which gives the private sector its accidental resilience. The diversity of business does not arise by top-down design but as an unintended consequence of competition: no one invested in home grocery delivery because it might prove handy in a pandemic, but because they were placing a side bet on how the future of shopping might play out.

From now on, we need to build the same resilience into everything, and spend less time over-optimising on the past. It’s one thing to teach computers to play chess; the rules haven’t changed in 300 years. In real life, as airlines found, the rules often change overnight.