Money And Robots And Men

[wonderpup]

No, that was Harlan Ellison.

According to Cameron the Director it was neither, apparently- he disputes Ellison's claim to have been ripped off. I must have been thinking of Blade Runner which was based on Dick's 'Do androids dream of electric sheep?.'

[wonderpup]

When discussing robots and AI most people have a tendency to anthropomorphise things. Most of these discussions, and this one in particular, seem to be trojan horse topics to provide a way for thinking in particular ways about human societies, not the robots of the thread title.

Not really a Trojan Horse- just an entirely transparent attempt to tease out an idea as to what function money actually has in human society by eliminating the human element and imagining a purely functional arrangement.

The fact that our robot civilization could happily mimic the entire 'capitalist' process from initial resource extraction to a final product without any need for a financial system is interesting in the sense that it undercuts the assumption that money is somehow essential to this process in an intrinsic way- clearly it's not.

Which suggests that the real function of money is not primarily to facilitate the production of goods and services but to facilitate their distribution, making money an entirely political and social thing, rather than something that is absolutely vital to the very ability to turn a raw material into an end product.

The meme that states that the market mechanism is the best and most efficient way to allocate resources is only true if one has internalized the premise that this resource allocation must take place in the context of a fractious and divided human race.

It would be more accurate to claim that the market is the least offensive way we have yet devised to distribute the spoils of what can still be an appallingly inefficient and environmentally destructive 'wealth creation' process.

The market is the best we can manage because we lack the maturity as species to allocate our resources in a more intelligent manner.

To risk another aphorism; money is the continuation of war by other means.

[scepticus]

Not really a Trojan Horse- just an entirely transparent attempt to tease out an idea as to what function money actually has in human society by eliminating the human element and imagining a purely functional arrangement.

The fact that our robot civilization could happily mimic the entire 'capitalist' process from initial resource extraction to a final product without any need for a financial system is interesting in the sense that it undercuts the assumption that money is somehow essential to this process in an intrinsic way- clearly it's not.

Well in my post I did posit that it would be necessary for the machines to lend and borrow rights to suck up the systems free energy at any given instant. That mechanism would need a means of accounting, and would involve credit and debt.

I think what it shows is that credit is the fundamental concept in our society and in the example machine one. In both cases a unit of account is needed, but not what we term as cash money.

As to what mechanisms and imperatives the machines here have for determining their energy needs at any given time I left to speculation, since clearly that's all it can be. If they are, as suggested, a society comprised of many distinct actors all of whom wish for self preservation and propagation then the degree to which they employ a market or they employ planned or altruistic transfers rests on the nature of the nash equilibrium that arises. In that sense they are no different to us.

[zugzwang]

Well in my post I did posit that it would be necessary for the machines to lend and borrow rights to suck up the systems free energy at any given instant. That mechanism would need a means of accounting, and would involve credit and debt.

I think what it shows is that credit is the fundamental concept in our society and in the example machine one. In both cases a unit of account is needed, but not what we term as cash money.

As to what mechanisms and imperatives the machines here have for determining their energy needs at any given time I left to speculation, since clearly that's all it can be. If they are, as suggested, a society comprised of many distinct actors all of whom wish for self preservation and propagation then the degree to which they employ a market or they employ planned or altruistic transfers rests on the nature of the nash equilibrium that arises. In that sense they are no different to us.

Assuming, of course, that a Nash equilibrium is computable - there's no reason why that should be the case - and that the time taken to attain equilibrium (exponential in the number of agents and their memory size) is not impractical. In real financial markets these twin constraints (finitistic computability and computational complexity) ensure that equilibrium is never observed, or even sought for. In that sense, at least, your hypothetical robot society is very different to us.

[wonderpup]

Well in my post I did posit that it would be necessary for the machines to lend and borrow rights to suck up the systems free energy at any given instant. That mechanism would need a means of accounting, and would involve credit and debt.

I think what it shows is that credit is the fundamental concept in our society and in the example machine one. In both cases a unit of account is needed, but not what we term as cash money.

As to what mechanisms and imperatives the machines here have for determining their energy needs at any given time I left to speculation, since clearly that's all it can be. If they are, as suggested, a society comprised of many distinct actors all of whom wish for self preservation and propagation then the degree to which they employ a market or they employ planned or altruistic transfers rests on the nature of the nash equilibrium that arises. In that sense they are no different to us.

The difference is that their 'credit' system would be driven by empirical data, while ours is driven by an abstract overlay of 'rights' and 'obligations' many of which are rooted in the hierarchical power structure. In our society claims on the energy of the system can be determined not by the optimal outcome for the system as a whole but by some completely arbitrary factor such as being born into the right family or being in a position to game the system to redirect it's resources to one's self- there is no real distinction made between activity that genuinely benefits the whole and activity that is detrimental to the whole.

For this reason money is in fact a poor tool for the allocation of resources in an intelligent way- it only seems optimal because we define the optimal as being the creation of bookable profit in accounting terms while carefully screening out such externalites as environmental degradation, resource depletion and absurdities such as simultaneous obesity epidemics in one region combined with mass malnutrition in another.

Whatever virtues can be claimed for our system even the most starry eyed enthusiast must surely admit that the intelligent allocation of resources is not one of it's strengths.

[scepticus]

The difference is that their 'credit' system would be driven by empirical data, while ours is driven by an abstract overlay of 'rights' and 'obligations' many of which are rooted in the hierarchical power structure. In our society claims on the energy of the system can be determined not by the optimal outcome for the system as a whole but by some completely arbitrary factor such as being born into the right family or being in a position to game the system to redirect it's resources to one's self- there is no real distinction made between activity that genuinely benefits the whole and activity that is detrimental to the whole.

For this reason money is in fact a poor tool for the allocation of resources in an intelligent way- it only seems optimal because we define the optimal as being the creation of bookable profit in accounting terms while carefully screening out such externalites as environmental degradation, resource depletion and absurdities such as simultaneous obesity epidemics in one region combined with mass malnutrition in another.

Whatever virtues can be claimed for our system even the most starry eyed enthusiast must surely admit that the intelligent allocation of resources is not one of it's strengths.

I don't think maximally efficient or 'fair' distribution of resources is compatible with a viable complex structure, whether machine or human. The unfairness based on past outcomes you refer to is actually part of the whole system's memory. If the system as a whole entirely disregards the past when allocating resources for the next step then it exhibits no memory. The whole point of system memory in a complex adaptive system is to retain records on past outcomes which can be used as a basis for predicting future outcomes.

In human society this manifests as stuff like inheritance and repeating dominance patterns. But don't forget, humans need some laws to live by and some enforcement of said laws, and without power hierarchies developing in such a way that they can persist at least long enough to enact said enforcement we don't have a system at all, we have chaos.

Which is not to say that the power structures should not evolve - they must do so for the system to persist over the long term and to grow where opportunities exist - merely that the time constant for change needs to be slower than the local rates of change in the natural environment that the system must react to and survive.

[scepticus]

Assuming, of course, that a Nash equilibrium is computable - there's no reason why that should be the case - and that the time taken to attain equilibrium (exponential in the number of agents and their memory size) is not impractical. In real financial markets these twin constraints (finitistic computability and computational complexity) ensure that equilibrium is never observed, or even sought for. In that sense, at least, your hypothetical robot society is very different to us.

I didn't mean to imply the machines compute a nash equilibrium, merely that a situation evolves which could be reasonably assumed to be a nash equilibrium or some hierarchy of the same from the viewpoint of some observer looking tino the system from the outside. Not sure how that is really different to the situation of the human economy?

Surely any system of interacting, selfish entities subject to real world physical constraints would qualify?

[zugzwang]

I didn't mean to imply the machines compute a nash equilibrium, merely that a situation evolves which could be reasonably assumed to be a nash equilibrium or some hierarchy of the same from the viewpoint of some observer looking tino the system from the outside. Not sure how that is really different to the situation of the human economy?

Surely any system of interacting, selfish entities subject to real world physical constraints would qualify?

Economists have half a dozen definitions of equilibrium. Most of them derived inconsistently from 19th century physics, and all of them wrong. Except under trivial circumstances Nash equilibria are intractable. Real financial markets are complex in ways that are not amenable to any kind of equilibrium analysis, even approximately. As Per Bak said, equilibrium is the condition of a glass of water at rest. Nash's failure to recognise the towering insights of Godel and Turing in his work should have given the game away; in fact his colleagues at RAND were generally dismissive during his time there and afterwards (von Neumann knew Godel personally, of course). Nash's post-hoc canonisation in the 80s and 90s from 2nd rate mathematician to neoclassical cornerstone was largely a political contrivance designed to reinforce the idea of sovereign markets and rational actors. Nash's optimisations make infinite demands on the computational resources of those who attempt to employ them, something that could have been established empirically, but wasn't - like so many false conjectures in modern economics. Adam Smith certainly didn't make the same mistake, he believed that moral sentiment was every bit as important as rational calculation.

[wonderpup]

I don't think maximally efficient or 'fair' distribution of resources is compatible with a viable complex structure, whether machine or human. The unfairness based on past outcomes you refer to is actually part of the whole system's memory. If the system as a whole entirely disregards the past when allocating resources for the next step then it exhibits no memory. The whole point of system memory in a complex adaptive system is to retain records on past outcomes which can be used as a basis for predicting future outcomes.

In human society this manifests as stuff like inheritance and repeating dominance patterns. But don't forget, humans need some laws to live by and some enforcement of said laws, and without power hierarchies developing in such a way that they can persist at least long enough to enact said enforcement we don't have a system at all, we have chaos.

Which is not to say that the power structures should not evolve - they must do so for the system to persist over the long term and to grow where opportunities exist - merely that the time constant for change needs to be slower than the local rates of change in the natural environment that the system must react to and survive.

Surely the point is that a system based on empirical data will be reacting in real time to changes in it's environment- whereas a system based on historical rights and obligations will not? Much of history is a story of societies- or more accurately their elites- that failed to respond in a timely manner to changes in their environment because that elite was determined to defend it's current position.

The problem with Adam Smith's invisible hand is that it works well to allocate resources according to local cultural understandings of supply, demand and ability to pay- but is completely blind to any wider concerns as to the viability of that culture as part of the larger ecosystem in which it operates. That invisible hand would cheerfully chop down every tree on the planet if the demand for their wood was present- each man with an axe acting in his own self interest of course.

The notion that money is a value free arbiter of resource allocation is pure fantasy-in reality it's something close to being the opposite-money is about resource allocation within a network of local values and social conventions- it's a means not ensure the most rational allocation of resources but the most advantageous allocation of resources toward whichever Elite group happens to be in control of the system.

You only have to look at the way in which the current Elites completely abandoned any adherence to the doctrine of 'rational markets' when their own wealth and power came under threat, at which point they intervened to ensure that the outcome was one more to their liking. When that invisible hand came knocking at their door they sent it away.