Great piece Kevin illustrating again that 1) UBI is a broad term with different meaning for different people. 2) The mathematics make the cost very difficult to justify in the AGI context.
I have a couple of pieces I am working on that discuss this. First, historically, automation has created more jobs than it destroyed. Given this, we could argue that fears of AI mass unemployment are overblown.
On the other hand, if AI does offer a “better than human” stand in for labor and cognitive tasks (assuming AI is also controllable) then the explosion of ideas creation will lead to incredible economic growth while also potentially driving down the value of human labor.
Both scenarios have positives, the former doesn’t require UBI, the latter creates incredible wealth where it might make UBI possible.
For me, the solution then is to begin replacing welfare programs with cash benefits now. That way, depending on which future path turns out to be correct, we can simply adjust the amount of cash distributed over time.
Thanks JK! a) Fully agree that automation has been net gain for jobs so far. I think it's plausible that this changes if and when we reach AGI with human/superhuman fluid intelligence but agree that best way to think about this is probabilistically. b) The idea of starting small and increasing the UBI if necessary sounds very reasonable.
honestly i'm surprised i've never heard UBI framed this way until now.there's a leap here that funding for UBI would come from an income tax, which i'm not wholly sold on. technology is deflationary, or such is the trope, and to what degree you accept that as accurate would seem to drive a lot of how you think about "post-labor" economy.
because if technology _is_ deflationary -- if technological growth decreases the costs of production -- then the real cost to providing the "base necessities" ought also to decrease in time, absent shifting definitions for what those "base necessities" are. "post labor" in the *extreme* means that labor is completely absent from production: that all production is automated. then where is the scarcity? where do you point and say "*this* is why the products of a post-labor economy have a cost to them (and hence, are inaccessible to those without means)"?
taking "post-labor" seriously, then ask "if it requires no labor to manifest a factory which would provide all the material food, shelter, and basic needs, then what's preventing each person without other means from doing so?". the intuitive answer is "actually it _does_ still require labor to do those things, it just isn't legible". but that's just a denial of "post labor", and the question should shift to focusing on that failure to allocate labor.
the answers to the above which accept "post labor" have to confront that the constraints on production *no longer have to do with the differences between individuals*. they're things like "you can't manifest a factory here because someone else owns the land required for it", or "someone else owns the patents", and other such things where you're dealing with property rights which were at best obtained during a period _before_ this "post-labor" economy.
perhaps a practical answer to the above is that even in a post-labor economy, there are inescapable physical constraints. there's still a limited supply of the material inputs to such an economy: the different ores needing to be mined from the earth, energy inputs from the Sun, and so forth. if a UBI exists in such an economy which fully automates production downstream of those physical inputs, a price system serves to allocate those physical inputs, indirectly, to the people downstream of all this. it doesn't require taxation of any kind. it's just a recognition that if we didn't enforce a price system, those material inputs would be allocated wastefully. and i'm not sure what more palatable form of allocation would exist there other than "everyone gets a roughly equal share of the inputs, which we distribute by means of some UBI".
anyway, i don't expect this conclusion to be met with cheers. but i'm not sure how to avoid it except by claiming that post-labor is a lie, or that increased unemployment isn't wholly technological (which to be honest, is far more believable to me).
Thanks for the extensive comment - appreciated! 1) Fully agree that UBI does not have to be financed by income tax and there are proposals based on other funding sources - I just wanted to show that this type of financing is not sustainable in a post-labor scenario. 2) On deflation: I agree that we could plausibly see a lot of stuff getting cheaper with caveats that a) progress will be very uneven by areas b) the cost for products will not be zero - there will still be capex/opex. c) deflation is not desirable from a monetary policy view (eg consumption & public debt) and it's kinda easy to avoid, so if it happens I just expect money printers to go brrr. (That could also be one unorthodox source of funding.) 3) I'm not 100% sure I correctly understand your post-labor vision. Essentially we declare resources (inputs) as commons? Or they are state-owned and that finances a UBI in money?
> I'm not 100% sure I correctly understand your post-labor vision. Essentially we declare resources (inputs) as commons? Or they are state-owned and that finances a UBI in money?
i think the bit to probe here is about how post-labor interacts with post-scarcity. if post-labor implies post-scarcity (a utopian world where the resources at our disposal are beyond what any individual could imagine using for themselves), then what need is there for any formal economy -- and therefore UBI? if post-labor doesn't mean post-scarcity, then how do the scarce things (the physical inputs) get allocated?
to borrow from the Georgist view, the economy consists of labor, capital, and "land" (lot parcels, but also here the ores, energy inputs, the things which exist without humans). labor is gone. what then of capital? most forms of capital we recognize are a product of labor: some people labor to build a factory and create that most recognizable form of capital. others labor to produce the patents and processes associated with this production. we formalize the ownership of this capital, to finance the labor which creates new capital, and apportion ownership of that new capital accordingly.
if we remove labor, that freezes the ownership of capital. what even is being financed when capital creates new capital without labor anywhere in that chain? if it's not post-scarcity, then what's being financed is the acquisition of physical resources. and it it's post labor, then what's being financed is merely the legal _entitlement_ to such resources. we're left with some finite resources, and a complex system ("capital") for rearranging those resources. a new human is born into this world, but without labor he has no clear ties to any of it. how do you integrate him? is he given an annual allotment of those still-scarce resources, which he then sells and exchanges for the downstream products he wants? or is he allocated a share of the capital, which he sells to the same effect? does the state tax the capital flows and distribute those proceeds to him? does it just print money and distribute that in the form of UBI? i suppose i do find the last two more believable, and "just print money" might actually have the least overhead.
Thanks! 1) I am not sure capital ownership would fully freeze. The firms or other entities that own all the AI / robots / datacenters / mines / grids etc will presumably not perform uniformly / shares can be sold etc so distribution can shift amongst humans and non-human agents might to start own more of the economy over time 2) fully agree that there is an intergenerational challenge and there are a couple of possible solutions from universal basic wealth to UBI
Great piece Kevin illustrating again that 1) UBI is a broad term with different meaning for different people. 2) The mathematics make the cost very difficult to justify in the AGI context.
I have a couple of pieces I am working on that discuss this. First, historically, automation has created more jobs than it destroyed. Given this, we could argue that fears of AI mass unemployment are overblown.
On the other hand, if AI does offer a “better than human” stand in for labor and cognitive tasks (assuming AI is also controllable) then the explosion of ideas creation will lead to incredible economic growth while also potentially driving down the value of human labor.
Both scenarios have positives, the former doesn’t require UBI, the latter creates incredible wealth where it might make UBI possible.
For me, the solution then is to begin replacing welfare programs with cash benefits now. That way, depending on which future path turns out to be correct, we can simply adjust the amount of cash distributed over time.
Thanks JK! a) Fully agree that automation has been net gain for jobs so far. I think it's plausible that this changes if and when we reach AGI with human/superhuman fluid intelligence but agree that best way to think about this is probabilistically. b) The idea of starting small and increasing the UBI if necessary sounds very reasonable.
honestly i'm surprised i've never heard UBI framed this way until now.there's a leap here that funding for UBI would come from an income tax, which i'm not wholly sold on. technology is deflationary, or such is the trope, and to what degree you accept that as accurate would seem to drive a lot of how you think about "post-labor" economy.
because if technology _is_ deflationary -- if technological growth decreases the costs of production -- then the real cost to providing the "base necessities" ought also to decrease in time, absent shifting definitions for what those "base necessities" are. "post labor" in the *extreme* means that labor is completely absent from production: that all production is automated. then where is the scarcity? where do you point and say "*this* is why the products of a post-labor economy have a cost to them (and hence, are inaccessible to those without means)"?
taking "post-labor" seriously, then ask "if it requires no labor to manifest a factory which would provide all the material food, shelter, and basic needs, then what's preventing each person without other means from doing so?". the intuitive answer is "actually it _does_ still require labor to do those things, it just isn't legible". but that's just a denial of "post labor", and the question should shift to focusing on that failure to allocate labor.
the answers to the above which accept "post labor" have to confront that the constraints on production *no longer have to do with the differences between individuals*. they're things like "you can't manifest a factory here because someone else owns the land required for it", or "someone else owns the patents", and other such things where you're dealing with property rights which were at best obtained during a period _before_ this "post-labor" economy.
perhaps a practical answer to the above is that even in a post-labor economy, there are inescapable physical constraints. there's still a limited supply of the material inputs to such an economy: the different ores needing to be mined from the earth, energy inputs from the Sun, and so forth. if a UBI exists in such an economy which fully automates production downstream of those physical inputs, a price system serves to allocate those physical inputs, indirectly, to the people downstream of all this. it doesn't require taxation of any kind. it's just a recognition that if we didn't enforce a price system, those material inputs would be allocated wastefully. and i'm not sure what more palatable form of allocation would exist there other than "everyone gets a roughly equal share of the inputs, which we distribute by means of some UBI".
anyway, i don't expect this conclusion to be met with cheers. but i'm not sure how to avoid it except by claiming that post-labor is a lie, or that increased unemployment isn't wholly technological (which to be honest, is far more believable to me).
Thanks for the extensive comment - appreciated! 1) Fully agree that UBI does not have to be financed by income tax and there are proposals based on other funding sources - I just wanted to show that this type of financing is not sustainable in a post-labor scenario. 2) On deflation: I agree that we could plausibly see a lot of stuff getting cheaper with caveats that a) progress will be very uneven by areas b) the cost for products will not be zero - there will still be capex/opex. c) deflation is not desirable from a monetary policy view (eg consumption & public debt) and it's kinda easy to avoid, so if it happens I just expect money printers to go brrr. (That could also be one unorthodox source of funding.) 3) I'm not 100% sure I correctly understand your post-labor vision. Essentially we declare resources (inputs) as commons? Or they are state-owned and that finances a UBI in money?
> I'm not 100% sure I correctly understand your post-labor vision. Essentially we declare resources (inputs) as commons? Or they are state-owned and that finances a UBI in money?
i think the bit to probe here is about how post-labor interacts with post-scarcity. if post-labor implies post-scarcity (a utopian world where the resources at our disposal are beyond what any individual could imagine using for themselves), then what need is there for any formal economy -- and therefore UBI? if post-labor doesn't mean post-scarcity, then how do the scarce things (the physical inputs) get allocated?
to borrow from the Georgist view, the economy consists of labor, capital, and "land" (lot parcels, but also here the ores, energy inputs, the things which exist without humans). labor is gone. what then of capital? most forms of capital we recognize are a product of labor: some people labor to build a factory and create that most recognizable form of capital. others labor to produce the patents and processes associated with this production. we formalize the ownership of this capital, to finance the labor which creates new capital, and apportion ownership of that new capital accordingly.
if we remove labor, that freezes the ownership of capital. what even is being financed when capital creates new capital without labor anywhere in that chain? if it's not post-scarcity, then what's being financed is the acquisition of physical resources. and it it's post labor, then what's being financed is merely the legal _entitlement_ to such resources. we're left with some finite resources, and a complex system ("capital") for rearranging those resources. a new human is born into this world, but without labor he has no clear ties to any of it. how do you integrate him? is he given an annual allotment of those still-scarce resources, which he then sells and exchanges for the downstream products he wants? or is he allocated a share of the capital, which he sells to the same effect? does the state tax the capital flows and distribute those proceeds to him? does it just print money and distribute that in the form of UBI? i suppose i do find the last two more believable, and "just print money" might actually have the least overhead.
Thanks! 1) I am not sure capital ownership would fully freeze. The firms or other entities that own all the AI / robots / datacenters / mines / grids etc will presumably not perform uniformly / shares can be sold etc so distribution can shift amongst humans and non-human agents might to start own more of the economy over time 2) fully agree that there is an intergenerational challenge and there are a couple of possible solutions from universal basic wealth to UBI