Capitalist strategy when capital doesn't need capitalists anymore
Many have wondered about the coming redundancy of labourers; few have thought about the coming redundancy of capitalists. Even good work in this area (e.g. Anton Leicht) often seems to implicitly assume that capitalists will be allowed to go along as is.
Say that capital has been automated when strategic management decisions are made by AI rather than by a human. One can automate capital in principle as easily as labour; indeed, it may even be easier to automate capital than many types of labour.
At some point, not automating capital will amount to leaving money on the table at best, and giving serious ground to the competition at worst. AI will exceed human skill in areas utilised in capital management. The capitalist who would prefer to manage their own money will face competitive pressures against it.
The automation of capital and the events that accompany it will create pressure on the state to seize and nationalise capital. Such pressure will not just come from ordinary people, but from political and other elites. We can anticipate that capitalists will respond strategically to try to avoid losing their assets.
On the surface, modulo any safeguards, automated capital can be seized by the state quite easily. It can be left managed by its prior artificial manager, or a substitute, with only the owners/residual claimants changed.
But why think it makes a difference? States could already seize capital. It’s no one thing, but the whole flavour of the situation is changed.
Suppose it is the distant future, the year 2000.
90% of workers have lost their jobs, as labour is generally automated. Consequently, the people and politicians alike are demanding the nationalisation of capital. What is motivating them, and what are they arguing?
Concern about the stability of democracy and the system of power. If capitalists have the power to withhold access to goods and services, but other democratic citizens no longer have the power to withhold their labour, the unevenness of the situation must feel dangerous to the future of democracy, and rightly so. Bringing capital under state control might seem like the only sustainable basis for continued democracy, or, to put it less positively, the existing power arrangement.
A desire to own the means of life by the democratic public, now that they can no longer sell their labour to subsist, and a desire by politicians and state functionaries to control critical resources. By owning, through the state, capital in a post-labour economy, the proletarians secure for themselves more direct access to the means of life, rather than relying on taxation.
Disgust at capital income earned without talent or significant ongoing sacrifice- what is sometimes called ‘coupon clipping’. Of course, capital income often takes this form- consider the absentee landlord. However, the role of AI in the process as the decision-maker makes it even more obvious.
Retaliation. The worker resents having lost their leverage, being left at the mercy of state welfare, due to the action of the capitalist, and not unreasonably so. The worker sees no particular reason why the capitalist shouldn’t be in a similar situation as well.
Failure of economic arguments. If the state can give capital over to a friendly managing AI, equivalent to the one currently managing it, arguments that economic ruin will follow nationalisation are harder to uphold. The game between ordinary people and the capitalists is now much more zero-sum. Likewise, the argument that capital ownership by the private sector incentivises people to take risks to become grand capitalists is also defeated.
Death of the individual merit argument. AI doing the work makes it implausible that profit represents the genius or hard work of the capitalist and is thus just desserts.
Failure of the individual freedom argument. With the vast majority of people effectively holding almost no economic power whatsoever, and class mobility gone, the idea that individual ownership = freedom will seem less plausible than ever- less plausible even than a society in which property has been abolished for 9/10ths of the population. Likewise, the death of the ‘aspirational’. Aspirational members of the public will no longer think that they might themselves become capitalists.
As AI grows far more intelligent than humans, it may become apparent that whoever holds it is setting themselves up for unassailable power. Whatever we may believe about the relation between the state and money in normal times, no one wants to lose power forever.
Of course, we cannot predict what will happen with any certainty. In Anti-Dühring, Engels contemplates the possibility that the state may be compelled to seize control of joint-stock companies, because they represent a naked mode of exploitation with no real, active contribution by the capitalist. For Engels, this would not be the end of capitalism or the commencement of communism; it would be the beginning of state-run capitalism. In the main, this has not happened. Intensifying the pressures he described might not do the trick either. The state might well leave capital alone. However, point 8 is critical- they are risking permanent marginalisation if they do.
It will be obvious to the capitalist in advance of any attempt that seizure is a possibility. The capitalist will adopt strategies to avoid generating the conditions of their own expropriation.
A) They may try to strike first and seize control of the state in a business coup.
B) They may try to set up failsafes, dead man’s switches, etc., that make the expropriation of their capital harder.
C) They may try to spread out the capital goods they own across many different locations under different jurisdictions.
D) They may try to guide the development of the underlying technology in such a way that methods for automating capital are not so quickly developed. Marc Anderssen has already fantasised that the last job that needs to be done by a human will be venture capitalist; it is not inconceivable that he and people like him might try to make this a reality.
E) They might aim to remain well on the good side of a handful of powerful states, and perhaps their citizens, who can enforce their property claims in other places [e.g. the United States- continuation of existing strategy]
F) They might seek military power for themselves.
G) They might try to escalate the speed of AI advancement to further increase their advantages, making seizure more difficult or impossible.
And G, of course, carries so many further risks in itself.
For those who want to create a future rich in the wonder of humanity generally, not just whichever psychos happened to be at the helm at the time more than human intelligence was created, it’s a conundrum.
Congratulations on the ACX shout-out.
> At some point, not automating capital will amount to leaving money on the table at best, and giving serious ground to the competition at worst. AI will exceed human skill in areas utilised in capital management. The capitalist who would prefer to manage their own money will face competitive pressures against it.
Replace AI with the technical intelligentsia, and this has largely already happened. There has not been much if any redistribution or disempowerment of capitalists as a result.
Individual capitalists doing their own investing still exist, of course, and there are even a handful of Warren Buffets, but the bulk of capital allocation is carried out by salaried employees - with very, very high salaries mind you, but still nothing remotely close to the incomes of their employers.