Results 1 to 20 of 147
Super Computers and a Planned Economy
Are there any studies done on this? I read that in the year 2000, it was predicted that a super computer would be built which has a computing power greater than the human race by the year 2013. I haven't heard any developments on this, so I will assume that this isnt near completion. It also brought to mind Hayek's criticism of a planned economy: it assumes that the intelligence of one man or agency is equivalent to the intelligence of all humans in the world in allocating resources. (Not that I agree with everything that Hayek says, but I'm sure a lot of policy makers do.)
+ YouTube Video
Will the development of a supercomputer like this make it possible for a planned economy to be obviously better than a market economy?
Last edited by smk; 8th May 2011 at 04:03. Reason: originally used inaccurate terminology for the idea I was expressing (used 'efficient' to describe the best possible system)
"The law of life should not be competitive or acquisitiveness but cooperation, the good of each contributing to the good of all." ~Jawaharlal Nehru
[FONT="Arial Black"]Inquilab Zindabad[/FONT]
This is an idea that has been around for awhile and occasionally finds its way here.
After having read several comments/manuscripts/etc... on the problem, I think even today's supercomputers can solve the economic calculation problem at least for a small country. I don't know whether this fact alone makes a planned economy inherently more efficient (much less just) than a market economy, but it does overcome one of the central theoretical obstacles to implementing socialism.
百花齐放
-----------------------------
la luz
de un Rojo Amanecer
anuncia ya
la vida que vendrá.
-Quilapayun
To address the video, why not a decentralized planned economy?
Sunt lacrimae rērum et mentem mortālia tangunt.
While a planned economy ran by a supercomputer or even a network would be magnificent the ultra-individualism instilled by the bourgeoisie would cause most people to react negatively.
------------------------------------
Hayek seems to be ignorant leftist theories and socialism considering he targets one idea that isn't held to be true by all socialist as critical to socialist ideology. Another thing that confuses me is how he treats it like a vacuum where all resources are distributed exactly 100% and then people are just expected to deal with it.
I think that Hayek is referring to the instantaneous ability to react to situations. I don't think that a decentralized planned economy without a supercomputer would solve this problem. however, a decentralized planned economy is obviously more favorable than a centralized planned economy.
"The law of life should not be competitive or acquisitiveness but cooperation, the good of each contributing to the good of all." ~Jawaharlal Nehru
[FONT="Arial Black"]Inquilab Zindabad[/FONT]
I don't quite understand the obsession with critiques of central planning written 50 or 100 years ago, whether by "free market" types or anarchists (who had raised many of the same points before, without getting the same attention in the mainstream media for some reason). None of them would have a snowball's chance in hell of getting hired by Google today.
(What I also don't understand is the annoying trend of linking to videos instead of text when the person in question has made the same claims in text.)
Paul Cockshott and Allin Cottrell have written extensively about that topic. Also take a look at their personal sites linked there.
Storing data and transmitting data and making calculations is getting easier and easier, hence the case for central planning is getting stronger. Cockshott's and Cottrell's book is now 2 decades old, the economic calculations they talk about probably can be done with the "super computers" you are sitting in front of right now.
It depends exactly what you're asking a planning computer to do - there are different types and levels of economic planning. If you're asking it to model the general demands for major productive resources - electricity, mineral commodities, transportation - then you can probably do that to some degree of accuracy on a discount laptop. If you're asking it to predict, perfectly (or even to the degree of accuracy as markets claim to achieve), the wants and needs of all persons, then it can't do that possibly, because there are no mechanisms by which it could get that information without humans directly inputing it. But why couldn't they? Almost everyone in many advanced capitalist countries has regular access to the internet. Making our general preferences clear could take a trivial amount of time compared to the time (and computing resources) spent on online shopping.
At any rate, no one thinks that every economic decision, every button press on a machine must be singularly and uniquely decided by one very specific entity, bureaucratic or digital. That's why we have workers. I think bourgeois economists making price-information arguments generally assume totally non-sensical levels of "planning centralization" that miss what the word "plan" means. No one is asking planning authorities to make all decisions for us. That doesn't make any sense.
No. Hayek is right in one sense but completely wrong in another - in assuming socialism must be a centrally planned economy encapsulated by the totally absurd abstraction of a "single society wide plan". Even if some supercomputer could coordinate all of society's inputs and outputs in this way via a vastly complicated form of linear programming, this only scratches the surface of the problem. Data collection and enforcement of the plan are much more formidable, and indeed insurmountable , problems to contend with
Recognising that socialism will NOT be a centrally planned economy in that sense, that it will have a feedback mechanism that can only exist when you have a polycentric or decentralised system, is the key to fundamentally and completely demolishing the whole Misesian economic calculation paradigm against a moneyless socialist economy. It is probably for this reason that anarcho-capitalists cling to a model of socialism that entails central planning in this sense. Intuitively I suspect they realise they have no answer to the argument that socialism will not - and could not - be centrally planned in that sense
I said a normal computer can do this.I agree that getting input that is close to reality is the tricky part, but I don't think that's insurmountable, unless meant in the sense of getting it perfectly right. In that sense many optimizing challenges are "impossible" that are nevertheless tackled with satisfying results in the real world. I call this critique of planning the Non-Travelling Emo Salesman Problem.
It's like saying that it's maybe possible to measure the direct work applied at the last step of producing everything, but impossible to measure the whole process. If the amount of work at the last step can be measured, we can go backwards from there to approximate the whole process. So the remaining question is how to measure the effort at that step, and when setting how much to give more for harder work and what the harder work is and what work is similarly stressful even if it is in completely different sectors, there's definitely a political dimension to it, that problem is not unique to planning.
I have found people who talk about "decentralized planning" to be rather vague about what they mean by that, I guess it's either bull or just central planning with some critique of some form of central planning nobody actually advocates.
replies to the hayekians
Here are some responses : http://www.dcs.gla.ac.uk/~wpc/reports/berlin.ppt
http://www.dcs.gla.ac.uk/~wpc/report...onearticle.pdf
orr this : http://www.dcs.gla.ac.uk/~wpc/reports/hayek/hayek.html
And a variety of other ones here: http://www.dcs.gla.ac.uk/~wpc/reports/index.html#econ
Whether or not a normal computer can do it or it requires a supercomputer, the computational aspect of central planning is the least of the problem
You can only say that if by "central planning" you mean something quite other than literal society wide planning which is what Mises and Hayek had in mind as an ideal type. That concept of central planning is unquestionably totally impractical - the idea of coordinating all society's inputs and outputs within a single mega matrix. Amongst other things the slightest perturbation in the real world will necessitate re-drafting of the plan in its entirely becuase of the interconnectedness of everything. And one thing you can predict is the unpredictable will happen. This alone disposes of the idea of the idea of central planning in the sense I am yalking about. But there are other problems - such as the problem of data collection for literally millions of different kinds of inputs and outputs which render the whole idea of central planning in this sense , impracticable
Well then we are obviously not talking about the same concept of central planning, are we. I am talking about the ideal type employed by the likes of Mises and co. What are you talking about?
I am not much of a fan of labour time accounting and I think you grossly simplify what is a much more complex problem. There is also a theoretiucal point about bygones being bygones. How useful is it to draw on data concerning labour already expended for the purposes of planning for the future
Well this may be the problem - that we are talking at cross purposes. My critique of central planning ir oriented towards the anarcho-caps definition of the same - namely, society wide planning , the idea of a single plan for the whole of soiciety. Decentralised or polycentric planning can take many forms but my own preference is for a multi-tiered approach with global, regional and local levels of planning with the bulk of planning decisions being made at the local level
If that was unquestionably totally impractical, you wouldn't feel the need to say that it was. At least these aristocrats have the excuse of being dead for some time, what's the excuse of the living for being ignorant of recent (and not so recent, really) developments?
Reminds me of a toddler reasoning they can't be seen because they have their eyes closed. Interconnectedness is true regardless of whether you have central planning. Central planning looks at interconnectedness, this is its strength.An implicit assumption in what critics of central planning have written seems to be central planning always meaning optimizing for the moment, so not being able to deal with quick developments. But of course it's possible to have a bit of redundancy in a central plan, for example I'd rather have a bit of overproduction in food items during normal times.Originally Posted by Kotze (from an essay)pretty much doable with a system of electronic payment and item tracking :POriginally Posted by robbo203Labour value estimates are to be used for deciding between different production processes for the same thing. Occasional cases where this returns very close results in their overall labour scores while having very different types of work in them, so that a change in politics giving a different multiplier for work classified as particularly arduous results in a different decision of what is considered efficient, doesn't prove it an unworkable concept. Under capitalism, cartels and a weird structure of taxes and subsidies make sensible cost calculation much more noisy.Originally Posted by robbo203
That's exactly what I referred to with Non-Travelling Emo Salesman. You don't need to know the most efficient route to travel and you don't need to know exact labour values for production, and in both cases having approximate solutions is very helpful.
Your oblique style of writing leaves me a little puzzled. I simply want to enquire how you define central planning and not be delivered a lecture on dead aristocrats
I am using the classical definition of central planning as society wide planning with a single plan to coordinate all society's inputs and outputs because this is the definition used in the discourse on the subject. The assumption made by anarcho caps is that socialism would not work for the very reason that it entails central planing in this sense. Dont believe me? Well, do some reading yourself and check it out. Try D R Steele's From Marx to Mises Post Capitalist Society and the Challenge if Economic Calculation . Steele is adamant that this version of central planning is or certainly was commonplace among socialists and that there is even prima facie evidence for this in the works of Marx's and Engels (particularly the latter's Socialism Utopian and Scientific). So its not quite as irrelevant as you seem to make out.
My take on it is not only is this conception of central planning wholly impracticable; it is also incompatible with socialism being necessarily top-down in orientation. I am essentially attacking the anarcho caps for assuming that socialism must be centrally planned in this way - as they do despite what you say - but if any who claim to be socialist likewise promote such conception then my attack would equally be directed at them as well. Whether that includes you I do not know. And that was why I was asking how you define central planning
Of course "interconnectedness is true regardless of whether you have central planning" but what of it? Central planning you say looks at interconnectedness and that is its strength. It may "look" at it but can it cope with it? I suggest not. Its the very complexity of modern day production - its interconnectedness - that completely rules out central planning as defined above. The gulf between the plan and the economic reality would widen dramatically because the errors that are certain to happen would have cumulative ramifications. But then again I do not what you mean by central planning so we might talking at cross purposes
The whole Central Planning vs Free Market question is to a large extent a false dichotomy. Planning is rather essential within capitalism, be it that it mostly happens on a corporation scale. Communists simply want to extent this planning to the social level, so production based on human need is a feasible objective.
I think, thus I disagree. | Chairperson of a Socialist Party branchMarxist Internet Archive | Communistisch Platform
Working class independence - Internationalism - Democracy
Educate - Agitate - Organise
This is a good point. If you look at the amount of economic activity that goes on in the "financial services" or "marketing" sectors, functions often analogous to planning functions in a capitalist society, huge surplus goes into it and it's extremely inefficient. I've heard estimates that marketing takes up about a fifth of the US economy, and financial services a similar amount. If you hear bourgeois economists explain the use and function of financial services, it's basically planning. If a capitalist economy spends a third of its resources (at least) on planning, they have a very difficult argument to make that a communist society wouldn't have the resources to do it, especially without the wastes and redunancies involved with capitalist planning, eg. Coke vs. Pepsi competition, securities fraud, engineering circumstances for profiting from currency speculation, etc.
Last edited by jake williams; 10th May 2011 at 11:36.
I totally agree. The modern, gigantic, monopolistic corporation plans everything it does down to the number of seconds its employees spend on breaks. And it does this with computers. The modern corporation plans production down to the last plastic wrapper, plans prices down to the last cent; it controls demand with advertising planned down to the last second of television time.
But all of this has been known at least from the 60s: Galbraith, Sweezey and others. Yet all you see in the media and, I suppose, in economics classrooms, is the magic of the marketplace.
I know media propaganda is extremely powerful, but how can an entire generation of Americans be completely ignorant about something as basic as monopolies?
Re technology required. We suggested super computers back 20 years ago. It is now possible to buy a computer for about £2000 that will do all we suggested in 1991.
I have just been doing benchmarks on the following machine
64 Gb RAM
2 sockets filled with Xeon E5620
Intel(R) Xeon(R) CPU E5620 @ 2.40GHz
64 bit - 4 Cores - 8 threads
Gulftown, based on Westmere Architecture - the 32nm shrink of Nehalem
L2 Cache: 4 × 256 KB
L3 Cache: 12Mb
The price 6 months ago was £2400
the benchmark problem was an N body planetary simulation for the sun, 4 major planets and 1018 Oort cloud objects. I found with my auto parallelising Vector Pascal compiler that this machine would deliver 7 Gigaflops on a real problem.
The planetary simulation problem is a maximally interconnected problem in that every planet affects, via its gravity, every other. Planning problems and the determination of labour values are characterised by sparse matrices and as such have a lower interconnection order.
I would be interested in practical experimental work any others on the list might try to see how rapidly they can get plan algorithms to converge on high end desktop machines today.
http://www.hpcinthecloud.com/hpcclou...bal_scale.html
This article might be of interests to people in this thread:
I don't think processing speed or any mechanical abilities would be the problem. The problem would be programing the damn thing to do something useful, like run the entire economy.
All very impressive, Im sure, but to do what exactly? Performing the calculations, solving the linear equations, is the least of the problem. In the real economy there are millions upon millions of different inputs and outputs. Not only do you have to collect the relevant data and ensure its accuracy but more importantly ensure the Plan is faithfully carried out to the letter. Is this remotely feasible? Of course not. Just one poor harvest in the American Mid West, say, will upset your calculations for the agricultural sector with ramifications that will multiply and accumulate and so undermine the usefulness of the Plan. This is just one of a virtually infinite number of ways in which best laid plans of mice and men can turn out to be hopeless inept. A complete waste of time and effort
This is to say nothing of what is implied by central planning - that production quotas are rigorously adhered to, that consumption levels are absolutely enforced through the strictest rationing. Is this compatible with the emancipatory nature of a communist society? I dont think so
I am not at all sure what you are getting at here but I presume your are referring to labour time accounting. Labour time accounting is sometimes associated with the proposal to institute labour vouchers although it is not quite the same thing. Marx talked generally of using labour time accounting as a tool for communist planning and presumably intended it to be used in higher or free access commmunism as well the lower stage of labour voucher communism.
I see little merit in labour time accounting and even less in labour vouchers (try imagine for one moment trying to "price" goods in terms of the labour time content. People who think its easy have no grasp of the problem). There are multiple problems that beset labour time accounting, not the least of which is determining the ratio of skilled to unskilled labour or indeed between different kinds of skilled labour. There is also the question of what exactly are you measuring - socially necessary labour time or actual time worked? And how useful is it anyway to use past labour embodied in machinery and so on for the purposes of future planning. How is the value of this past or dead labour distributed between products produced by living labour utilising this peice of machinery? Or the electricity used to power it? And so on and so forth
There are other more flexible planning tools available to a communist society that can be brought into play without having to go down the dead end road of central planning in its classic sense
