Technologies that aren’t making it big

The twittoverse has been abuzz over Neal Stephenson’s article on rockets and innovation. Rockets are our only means to reach space and Stephenson’s claiming that we’re locked in this state due to path dependence, citing a list of political contingencies that provided the research money to get rockets working so we could threaten to drop nukes on those godless Communists/Capitalist running dogs (delete as appropriate). Basically, it’s political determinism rather than technological determinism.

Alternatively, there’s this article from Newberry which usefully points out where Stephenson is factually incorrect (i.e. just about everywhere) then misses the point completely by waffling on about turbines. It’s also factually incorrect, badly edited, and occasionally incoherent, as pointed out by some of the smarter people in this debate. Still, buried somewhere in that article is the idea of closed technological loops – families of technologies that support and sustain each other and thus grow symbiotically and rapidly. The classic example is iron & coal & steam, which together formed the basis of the industrial revolution. It’s like rock, paper, scissors, only everyone wins. Coal refines iron ore to iron, iron and coal make steam engines, steam engines power coal mining and factories, steam engined iron railways move iron ore, coal and factory goods.

In response to this discussion, I’ll put forward two ideas of my own:

Firstly, we know very little about how innovations succeed. We know how innovation works as a process – intellectual freedom for researchers, strong links between research and business/end users, lots of bright people with experience, and a balance between state and private funding so that at each stage of development there’s money available. We’ve had enough nations take this approach, for long enough that on aggregate, this system works. What we barely know is how to predict which innovations will work. If we did, then venture capital investments wouldn’t face such poor odds.

Secondly, what seems to make individual innovations work is a complicated mix of many social factors and fundamental technological difficulty. The social factors include markets, regulations, social needs, the impacts of other technologies, the rate of development of other technologies, and the sheer resources that society sees fit to put towards a given innovation. Technological difficulty is what you don’t know that you don’t know. It wouldn’t be innovation if you did.

So, here’s some technologies that are not going anywhere in a hurry and haven’t been for quite some time:
1) Fusion power – It’s just technologically hard. Ask an researcher and they’ll tell you it’ll be working in about thirty years. “Thirty years” is what people say when they don’t want to say “no idea mate, but please keep giving us the cash”.
2) Satellite-based mobile telephony – Perfectly feasible, technologically, but land-based approaches started earning money faster, kicking off sufficiently good service that satellite-based approaches were only left with a tiny niche.
3) Space-based solar power – Again, perfectly feasible, but land-based is cheaper and will always be so. Now that land-based approaches are growing at 30% per year, providing an income stream for competitive companies to do more research into driving down the cost of land-based solar power. (Yes, some of that research will help lower the cost of space-based solar power, but it will lower the cost of land-based solar power more rapidly than space-based. Hence an ongoing increase in the cost advantage of land-based.)
4) Reusable launch vehicles – Meh. Rockets are cheap enough, in comparison to their payloads, and the demand for launches seems inelastic. Dropping the price of launches by half isn’t going to double the number of launches.
5) Segways – “Cities will be built around this device.” Err… no, we’ve already built cities. They’re for cars (and occasionally for pedestrians).

18 thoughts on “Technologies that aren’t making it big

  1. What do you expect from those people – they’re not scientists, just a sci-fi author and a blogger, neither of whom are likely to produce accuracy.

    ….

    Satellite phones are limited by physics. Any (RF) communication system is limited in bandwidth by the amount of spectrum allocated to it. Putting the signal in a closed medium, like wire or fibre, mostly avoids this limitation, as you can use all the practical spectrum in that medium for a point-to-point circuit. So a communications satellite uses all its allocated bandwidth to deliver a very limited amount of capacity. Hence it can’t support many users within it’s bandwidth allocation, and is always going to be an expensive niche technology.

    1. They’re also not innovation policy wonks, although I’m not sure if that’s a good thing or a bad thing.

      Anyway, there’s two answers to the sat phones question. The first is that narrow beams and steerable antenna mean you can fit plenty of point-to-point circuits into a satellite. The second is that Iridium never got the chance to see how many “plenty” actually is, and whether that “plenty” is enough to support the research base to turn the 1990s version of plenty into the 2010s version of plenty.

      Basically I’m arguing that all technological constraints exist within a social and economic context. That context may provide research funds to throw at the problem. Those research funds may alleviate the constraint. They may not, yet. The context may change rendering the entire question moot. It’s all contingent (which is my way of saying that this stuff makes my head hurt).

      1. Sat phones solved a problem no-one had, and didn’t solve the problems they did. In the meantime, the problems we did have got solved without direct satellite involvement. Now we might care more about the space sat phones existed to solve, but we’ll do it by abstraction away from needing a special phone *just* to do that one thing.

        There’s a certain amount of “but it should be better” and it just isn’t. It only looks better on paper and in a lab, and where getting everyone to change is easy.

    1. Yes, but if it’s such a great idea then why hasn’t it been done already?

      (Eventually I plan to get paid a great deal of money to work on exactly this question.)

      1. Because the time travel device that the AI invents only allows you to go back in time to the point where the time machine was built. And *that* will be the singularity.

        1. That’s a bit of a crap singularity then. Can’t the Omega Point just simulate the existing universe without that limitation, then chuck us all into that simulation instead? Coz waiting for time machines to be invented is a tad dull.

          1. Well it certainly could, but it all depends on what nested level of time you’re referring to. Whether it’s relative to our existing selves or those being simulated! Of course, we won’t find out until we reach the Omega point of our current nesting…

          2. Yeah, but given the infinite number of potential simulations where unconstrained time travel is supported, the fact that we’re in a universe where it isn’t supported implies we’re not in a simulation. Thus the Singularity is impossible. QED and yah boo sucks!

            * – yes, I know this is arguing from gibberish, but that’s what happens when you let philosophers play with infinities.

          3. So, after several thousand years of Western thought, we can conclude that we’re not sure, really, about anything, and that more research is needed?

  2. You might find “What Technology Wants” by Kevin Kelly interesting… all sorts of stuff, including making an odd case for a kind of evolutionary determinism/inevitability for technological innovation.

    1. Or I might find it makes me want to put my head in a drawer and repeatedly slam it closed, like most things that Kevin Kelly writes.

      Actually, I might be better off putting his head in the drawer while slamming it and yelling “drawers are technology, so drawers are good, sacrifice your feeble head-meats to the glory of the drawer!”

      (Ooo… I am wearing my grumpy pants this morning.)

  3. I wrote a response to the Stephenson article on slashdot, that I’ll gloss here: he points out all the reasons everyone is sticking with rockets: enormous development costs that have already been spent, so people working now have the benefit of all this research; knowing that something can work before you start means you’re much more willing to invest; you can purchase expertise for rockets — who can help you if your launch loop keeps breaking?; and there’s a lot of political and economic pressure to keep pouring money into established rocket companies and their employees.
    More fundamentally, this is a problem with evolutionary processes: a system climbs the local maximum and once it’s there it’s stuck there, even if there are other taller maxima, because in order to get there you have to descend from where you’re at, and that’s not economically competitive. So the only ways to get there are if something happens that alters the entire landscape, making the local maximum no longer actually a maximum — which would in this case require changing physics or making some radical discovery — or if someone is willing to fund a $4 trillion move down and across to another maximum, and nobody is for the above-mentioned reasons.
    There’s no mystery to that: it’s just how businesses work, and since businesses stand to make money with the status quo, and whole new businesses would have to come into being for a different system, the existing businesses spend lots of money preserving the status quo.

  4. 4) Reusable launch vehicles – Meh. Rockets are cheap enough, in comparison to their payloads, and the demand for launches seems inelastic.

    Wait. Aren’t the payloads engineered to be super-reliable BECAUSE launch vehicles are expensive? Conceivably, much cheaper launches might evolve much cheaper payloads. How sure are you that the demand for launches is inelastic?

Leave a Reply

Your e-mail address will not be published. Required fields are marked *