> There are a factions in the strong AI community whose members espouse the idea that a planned economy would become feasible given greater computational capacity and better software.
The calculation problem with central planning isn't really a calculation problem, it's a missing input problem. You can't optimize utility without a good input on utility.
Capitalism has significant problems by being forced to treat money as equivalent to utility, when it manifestly suffers decreasing marginal utility like everything else. This means people with access to more of it are likely to give up more for the same utility, which means capitalism overweights the preferences of the rich, and this gets worse the more effective capitalism is at optimizing by the one utility measure it has. But centrally-planned systems have even worse utility inputs, which no measure of computational power can correct; it's a GIGO problem.
The experiences derived from that basic problem are (one of several reasons) why market socialism is much more common in socialist circles than state socialism these days (market socialism has been competing with state socialism—which many market socialists have referred to as state capitalism on the view that, in it, the state becomes one big capitalist enterprise—the whole time, but the Bolsheviks managed to fix most popular attention on their particular form of state socialism, and capitalists found it convenient to ignore other socialist viewpoints in their attempts to demonize socialism.)
>The calculation problem with central planning isn't really a calculation problem, it's a missing input problem. You can't optimize utility without a good input on utility.
Historically that has been true, but couldn't technology help? Like, issuing everyone a smartphone and having them register all demand and make all purchases through it. The rest is a problem of keeping capital from aggregating at the top.
> Historically that has been true, but couldn't technology help?
No, it's a fundamental structure problem, not a technology (aside from the sense in which an economic system is itself a social technology) problem.
> Like, issuing everyone a smartphone and having them register all demand and make all purchases through it.
From a utility signalling perspective, that's not significantly different in the information it provides than just retaining currency for consumer purchases, which state socialism generally has.
Or shouldn't be impossible to devise a robust decision and optimization algorithm against missing, wrong or inaccurate input.
That said, linear programming is not it and the problem is extremely slippery.
The trouble with governments as it is is that they make decisions In an extremely non-robust way...
The robust way is supposed to work like a many-worlds predictor- corrector pattern and that's just one way. It is an order of magnitude or two more computationally intensive. (Think robust weather predictions and modelling.)
I think it would pretty much require taking humans out of the decision-making process as much as possible, and maximal transparency where they are involved. Humans provide input as far as what they want and what they are willing to pay for it, and everything else is algorithms.
Apart from dragonwriter's point, that you're just duplicating the signaling that money currently does: You want to let some central authority watch every purchase I ever make, just so that they can try to plan the economy better? Um, hello? Privacy concerns, anyone?
If you don't need your ${purchase you'd like to keep private} to be optimally produced, you could register demand for a facility that allows you to select from an assortment of goods effectively anonymously, i.e. a shop. The demand would then be tracked on the aggregate level. Of course this is assuming that the selection includes what you want, otherwise you will have to be more open about your desires to have them satisfied.
That's a separate issue, but I imagine there's probably a way to collect the demand while keeping the individual anonymous. Besides, in the world of capitalism we have the same concern- Amazon, Google, my bank, and Verizon could all probably tell me things about myself and my habits that even I don't know.
The calculation problem with central planning isn't really a calculation problem, it's a missing input problem. You can't optimize utility without a good input on utility.
Capitalism has significant problems by being forced to treat money as equivalent to utility, when it manifestly suffers decreasing marginal utility like everything else. This means people with access to more of it are likely to give up more for the same utility, which means capitalism overweights the preferences of the rich, and this gets worse the more effective capitalism is at optimizing by the one utility measure it has. But centrally-planned systems have even worse utility inputs, which no measure of computational power can correct; it's a GIGO problem.
The experiences derived from that basic problem are (one of several reasons) why market socialism is much more common in socialist circles than state socialism these days (market socialism has been competing with state socialism—which many market socialists have referred to as state capitalism on the view that, in it, the state becomes one big capitalist enterprise—the whole time, but the Bolsheviks managed to fix most popular attention on their particular form of state socialism, and capitalists found it convenient to ignore other socialist viewpoints in their attempts to demonize socialism.)