The apotheosis of wealth building is a Von Neumann machine. We build one, it goes off and scours for the resources to build another, which it does, the two of them then scouring for the resources to build four, which they do, we end up with billions of machines with only the human labour to build the first one. Sure, there are practical difficulties here – like, well, how do we stop them scouring everything to build damn machines? But as a thought experiment about human wealth they’re great.
For, how much of anything that we can consume is ultimately limited by the human labour that has to go into its production. If we get to consume more per hour of human labour in production then we’re richer. Either we can consume the same amount and have more leisure or we can consume more for the same labour input. This is just another formulation of Paul Krugman’s productivity isn’t everything but in the long run it’s pretty much everything. If human labour is more productive then we’re richer.
So, this story of robots building robots is fun:
Swiss robotics company ABB has revealed that it’s spending $150 million to build an advanced robotics factory in Shanghai — one that will use robots to build robots. The company will rely on its YuMi single-arm robots, which it once used to conduct an orchestra, for small parts assembly. It also plans to make extensive use”of its SafeMove2 software in the facility, which it says will allow its YuMi models and other automated machines to safely work in close proximity with human employees.
We’re to use robots to build robots. We’re getting closer to that manufacturing of the future that employs a man and a dog:
Robots will make robots at a new ABB (ABBN.S) factory in China, which the Swiss engineering group said on Saturday it plans to build for $150 million in Shanghai as it defends its place as the country’s largest maker of industrial robots.
The factory, located near ABB’s China robotics campus, is due to be operating by the end of 2020 and will produce robots for China as well as for export elsewhere in Asia. China is ABB’s No. 2 market after the United States.
The man is there to feed the dog, the dog is there to bite the man if he tries to touch the robots.
Some think this would be a disaster, for where will all the jobs be if the machines do all the work? Doing things that machines can’t do, obviously enough. And if there’s nothing the machines can’t do then everything will be produced by machines. Which means we’ve no problem, for if everything is being produced then everything can be consumed and no human has to break a sweat making the stuff. We get everything without having to work – Marx’s nirvana that allows true communism brought to you, as predicted, by capitalist technological advance.
However, using robots to build robots isn’t quite our end game of Von Neumann machines. That would be when we use robots to build the robots which go repair the robots who mine the iron out of which we build the robots who build robots. Carry that on through however many iterations you desire. We’re not there but we’re getting closer. The burden of human labour is getting lighter and we’re all getting richer precisely because it is.
The language we use matters - it provides clarity to our own thoughts and enables…
It is now generally acknowledged that the structure of the NHS needs to be overhauled…
In the film Apollo 13, a loss of oxygen causes the crew to start inadvertently…
There's an idea out there which seems intuitive but then so many ideas do seem…
When we think about the darkly opaque goals of modern central bankers as they relate…
As the papers recently filled with the distressing images of desperate souls looking to escape…
View Comments
I'm going to try to think this through. Ok everything produced by robots.
Usually from there there it's hardwired straight to "no jobs!". And then: "no job, no money". "No money: i can't pay for anything so i don't get any of whatever the robots produce." That's the stonewall that's difficult to clamber over. For me its the re-valuation of everything that's difficult to take account of. Stuff that we think is expensive becomes cheap, stuff we think is cheap becomes (relatively) expensive. Stuff we don't even value or doesn't come into the cash economy will probably make an appearance as people move to niches where you still have to be human to do.
Maybe there will be people who charge for a human face to face conversation, oh sorry counselling already exists, but it could expand into new areas perhaps people would do it just for fun, viz the monty python sketch paying a person for an argument. Turned completely on its head you may have to pay for things that previously people were paid professionally to perform. Say paying a group of people to be taught French by you, because you like teaching French, the others would rather not but if you're going to pay them then well that's different.
Tim writes: "..., for where will all the jobs be if the machines do all the work? Doing things that machines can’t do, obviously enough."
The current and proposed circumstances (of tools and their use) have existed for millennia. Hammers were and are used to help humans build (other) hammers. Wheels were and are used to transport (and otherwise rotate) raw and partially processed materials - to build wheels.
The important step that has not yet occurred is for a machine to conceive all the machines needed to replace all human labour (including the labour of conceiving machines). This in sufficient detail to build one in practice. And don't come back claiming 'Turing machine' - that lacks the detail and the practice, though it does represent an important step: either on the way, or perhaps on the way to nowhere useful; and to many things more useful (now or soon) but less 'exciting'.
I repeat: the important step that has not yet occurred is for a machine to conceive all the machines needed to replace all human labour (including the labour of conceiving machines - all the machines needed to replace all human labour - including the labour of conceiving machines - all the machines needed to replace all human labour - ...).
But first we need another important step that has not yet occurred. That for any human to conceive (in sufficient detail) and then build one single machine of the type needed to replace all human labour. Because that will surely (have to) happen before we automate the process - no matter how many advanced tools (machines) we build to help us step along the way.
Some (even much) fiction is fascinatingly interesting; no fiction is fact. This applies as much (even more) to science fiction as it applies to any other sort.
Best regards
But, seemingly of course, a human conceiving of and building a machine that can conceive and build all human-conceivable-and-buildable machines will become out of date as soon as it is built. This because a machine more complicated than any previous one has now been conceived and built.
This reminds me a bit of a discussion on the blog of the Adam Smith Institute back on Fri 18 Dec 2015, concerning everyone eventually being paid the same wage. Then I wrote: "The lack of scarcity of skills implies no technological or other development with time, for which there would (initially at least) be a scarcity of the particular skill or skills to realise that development. Thus, this Utopia thingy is actually never achieved, because there is a better Utopia upcoming (as there is indeed a better vacuum cleaner upcoming - despite the wonderful developments to date). While some people might define a Utopia as adequate for them, they are surely pre-judging and constraining the Utopias to be imagined, to say nothing of the actualities to be realised, by future generations - and thereby these Utopian believers are holding back progress."
Best regards
aside: amusingly I had to prove that I wasn't a robot in order to log in and comment.
Something that gets missed is why we use machines, technology, robots, whatever. And the answer is that we do it because the capital cost of the machine is less than the wages of the people who would otherwise have to do the job capitalised over the expected lifespan of the machine. To say that once the machine is made by other machines then the capital cost will fall to zero is to make a mistake because there will be finite non-labour inputs into making the machine.
The informed reader will probably be able to add to these, but here are some finite non-labour inputs to start with:
1. material - even if the robots are doing all the mining then the permits to dig it out the ground aren't going to suddenly become free, in fact the lower the cost of production the higher the natural resource rent will be.
2. space - there's a finite amount of land, especially in the ideal places to build factories so there will be a natural bid-up on use of it. Nobody has successfully suggested how in the communist utopia, everyone can have a penthouse overlooking Hyde Park, even if scarcity of materials and labour had been 'solved'.*
3. energy - and combine that with space to limit transportation and infrastructure networks, too.
4. RF spectrum - lots of things talking to each other all the time, even with the most efficient use, there is a limit to the amount of information that can be sent over the useful parts of the RF spectrum.
I'm convinced that there will always be a market for blokes in sheds making one of something. Even when the machines are running everything, there'll always be something that is needed that is just beyond the capabilities of a generic thing-maker, and it isn't worth the time on the specialist machines to set up a production line to make one widget, so the head robot will get its secretary robot to ring Brian in Wolverhampton.
*interestingly, as society has become richer, people are willing to pay higher and higher premiums to live in a desirable place. In the 1970s a flat in Kings Cross would have been worth about 2 terraced houses in Bradford, now it'll be worth 15 or 20 of them.