Every now and then someone who has never developed anything new and taken it to market successfully comes along and looks at a technology that is in the early stages of development. Rockets, semiconductors, electricity, solar power, and so on.
They are inevitably horrified by the number of problems one would have to solve for the technology to succeed, the enormous investment required to solve those problems, the high probability of failure--and therefore the high probability of that often-public investment going to waste--and the sheer scale that a working solution would require.
Even experienced people can forget that what their personal brain finds staggering is not determinative of reality. The Director of the Office of Scientific Research and Development in the US, Vannevar Bush, was famously opposed to the development of rocketry as he believed it would require implausibly large and powerful rockets, comparable in size to navel warcraft, and he personally found it absurd that rockets on such a scale could possibly be practical... ever.
Today, it's difficult to search for his specific quote, because so many enormous rockets have been developed and launched since.
In the early 1960s Canadian physicist George Ewan invented the "lithium drifted germanium detector" which was the workhorse of gamma-ray spectroscopy for several decades. The lithium drifting process compensated for defects and impurities in a large germanium diode, and it was considered absurd that we would ever be able to refine materials to the point where there were essentially zero impurities. Except that within twenty years we did, and today the Ge(Li) detector has been entirely replace by the HPGe (hyper-pure germanium) detector.
When I was a student, a semiconductor technology based on gallium arsenide rather than silicon was considered to be the Next Big Thing in micro-electronics because silicon was "inherently limited" to under about 100 MHz speeds. The processor in your cell phone runs at over 1000 MHz. It is based on silicon.
And. So. On.
Every technological innovation has been considered absurd, unimaginable, not worth the risk (by those not taking any risk) and requiring implausibly extreme scales... before they were achieved.
Afterwards it was obvious to everyone all along that of course it would work.
Speaking as someone who has spent much of his career at the interface between lab and market, I can say with some confidence that anyone who claims to be able to do a better-than-random job at predicting which technological innovations will succeed and which will fail had better be willing to show their past predictions compared with their actual outcomes. Without that track record to demonstrate that the techniques they are using to predict yields results that are better than random, they should not be taken seriously. Why would they be?
After all, while their predictive techniques may seem reasonable, we know that "what seems reasonable" is a terrible guide to what actually works. You know what seems reasonable? Bloodletting. Prayer. Henry Ford said if he'd asked his customers what they wanted they would have said, "A faster horse." A faster horse is reasonable: people have been cross-breeding for improved livestock for ten thousand years. Taking just one and a half hours to build an entire automobile is not reasonable.
Why do technological prognostications continue to fall short, and why do people still believe them?
The latter is fairly clear: to do anything other than "follow your gut" is really hard. It takes a decade or more of highly specialized training and consistent, rigorous, practice, to learn to change your mind based on facts that point to conclusions that don't "seem reasonable." Most people think they can do this, but then, one in thirty women and one in ten men think they could win a point off Serena Williams, so it's possible that our self-assessed abilities don't always entirely reflect reality.
It follows from this that when the success of a new technology requires us to follow facts--like the success of many past technologies that were widely considered "obviously absurd" even by educated observers--people will for the most part dismiss the facts--often using quite creative special pleading--and follow their gut. Maybe one day we'll have a technology that makes it easier for people to be better Bayesians, but despite advances in education that seems absurdly unimaginable right now, because...
The reason why technical prognostications are so often so badly wrong is that they require us to imagine solutions to problems that we don't have solutions to: if we did have solutions to them we wouldn't even be having the conversation, because the technology in question would already exist.
This is a difference in attitude more than anything else. People committed to developing a new technology know that pointing out problems to be solved is just mapping the territory they have to explore. A list of problems is just the beginnings of a project plan.
People with different motivations--motivations that are frequently obscure, at least to me--often seem to think that pointing out the existence of problems that do not yet have a solution is something more than that... as if they were expecting a technology that does not yet exist to consist of nothing but a collection of easily solved problems with solutions that only require incremental developments based on current technologies.
But capitalism eats problems like that for breakfast. Their average lifetime is about a week. It follows from this that the only kinds of problems we face when developing something genuinely new are the ones that are absurdly hard and require major, non-incremental, developments, which no one can imagine how to solve.
If anyone could, they would have been solved already.
So pointing out that a new technology requires unimaginable solutions--as this article on lab-grown meat does--is uninteresting.
Does this mean everything will always work out? Of course not. Some things are genuinely beyond us. But we discover that by trying to do them, not by giving up because they look absurdly difficult.
This made me look at articles on things Steve Jobs said 25 years ago… interesting!