We'll need universal basic income – Geoffrey Hinton
bbc.com
bbc.com
I think however, that UBI may be worse than the Roman bread doles. Those were after all only food, but money are every kind of resource. There was also still a need for the client class, since they were in fact still a source of soldiers of the empire.
The change that UBI makes in modern times is that it changes people who are important and who co-operate in order to make things and get them to others are turned into just a cost which never produces anything. Consequently it'll always be appealing to get rid of them.
The only way ordinary people can survive is by having power, and UBI is not enough. If you need UBI you should probably have already had a revolution and seized control because the rulers can always end it.
Consequently I don't see how UBI can be a solution. It must be something else. I believe a reduction in working hours could work for a while, at least a decade, but won't be enough in the long run and I'm far from certain that there is a solution.
Another problem we've had here in Sweden is that people try to fund these things by taxing successful workers instead of capital owners etc., so it's a rebalancing within the working class while the genuinely rich are unaffected. Because any rebalancing of tax from ordinary people to the rich will cause inflation and reduced investment due to the fact that ordinary people spend a larger fraction of their incomes on consumer products it'll always be appealing to make this purely a rebalancing between workers.
Thus there isn't just one failure mode, but two.
The whole premise here is incorrect. Every time UBI is tried, this scenario does not happen. You can find study after study that finds the same thing. https://www.vox.com/future-perfect/2020/2/19/21112570/univer...
> Economists found that it doesn’t make them work less. It does lead to improved education and mental health, and decreased addiction and crime.
Are you aware of any UBI experiments where the recipients are confident that both:
1) the amount will cover their basic needs, and
2) the scheme will last forever?
The reason I'm asking is that if either of these conditions were not met, the recipient would be crazy to stop working.
If someone were to commit to giving you $10k/month forever, increasing with inflation, you might quit your job. If it were only $500/month, you'd probably keep working. If it were $10k/month, but the payments might stop after a year, you'd probably keep working.
(Not 'you' specifically of course. I know nothing of your personal situation.)
There is a tendency among many people to give precedence to a scientific study over pure reasoning even when the degree of relevance of the scientific study is tiny. It’s like schools don’t teach how to determine if a study is applicable and so people often do it very clumsily. This is a dangerous error.
I'm not saying I'm 100% certain it is invariant. Just that we should stop repeating the tropes that so far haven't been shown true.
I think pro UBI people often straw man anti UBI arguments as a disguised belief that Group X is lazy. I think pro UBI people do this because they are morally triggered by discriminatory work ethic beliefs (fair enough). I think that people who cite these studies might see the studies as ruling out the straw man, which makes them feel fuzzy self righteous feelings, but to me the studies do not even remotely address what I see as the actual risks of at scale UBI deployment.
If you work in a team programming something or similar, then you are co-operating to produce a product, and if you've got a civilised company you'll all end up friends. You've had a common goal and you've worked towards it together in an honest way. Furthermore, you are necessary to society. Even if the political and economic elite are not in on this work, and are not participating in it in an honest way, the success of your work is something they need, thus they need you.
But here the reason why we're instituting UBI is that people are no longer necessary to production.
This is similar to something about the resource curse. When countries need their workers for wealth they tend to need to ensure that their workers have good conditions, but if they can get away with just extracting oil or natural resources, then the good conditions disappear. AI will reduce the need for workers to the point of UBI, so it will become more like the extractive work in these states with less good conditions.
So my objection isn't to UBI as such, it's that I think UBI doesn't go far enough. Having thought some more since yesterday evening I think it may even be so that for the situation to be stable the people on UBI need to basically rule the country full-time. A mega-parliament with very great power. So as society moves away from work ordinary people need to move away from work and towards ruling.
https://www.utpjournals.press/doi/pdf/10.3138/cpp.37.3.283
All in about 4 years.
The problem is not much that the rulers can end it, but that they'll continue it in perpetuity. Or if you wish, that they will only selectively end it, or threaten to end it, for some - as a punishment of disobedience.
UBI's allure from the point of view of those in power is it turns citizens into dependents. So people not working, and not having power, thus turning into "clients", is kind the whole point.
So, yeah, the version of UBI that doesn't have that drawback, is UBI in a society where the masses have seized the control and are actively exercizing it too (not eager to give it back).
We talk about tax and money like they're natural theories, when we could easily pursue compatible, interoperable regimes that meet the UBI design brief. If price is value then change in price is a proxy for externalities, so tax transactions based on change in instantaneous price. The power to do so is in divide and conquer, between rentiers who value stability and those who cultivate volatility. The Not-In-My-Back-Office corporate leadership creates gigahertz iPads but never seems to buy their way out of overnight transaction clearing and filing labyrinthine returns for some reason.
I hate being livestock in the "incentive" farm. We almost all do. And while I appreciate insights about the Roman Empire as much as any white male...it's less the done thing these days. :)
The capital will just flee to other places if they do that. In fact it already does.
But I agree, UBI will never work. It will just generate inflation and drive prices up, like the Biden Administration's America Rescue Plan did by giving away $1400 checks to people.
https://groundworkcollaborative.org/wp-content/uploads/2024/...
A UBI would go to everyone, but the benefits would accrue largely to low-income people, hence my focus above.
https://www.baen.com/Chapters/9781481483063/9781481483063___...
Of course you don't need to agree with the conclusion!
> Critics say it would be extremely costly and divert funding away from public services, while not necessarily helping to alleviate poverty.
I'm not sure why the reporter wrote this. Have there been any failed UBI studies, ever? (https://www.vox.com/future-perfect/2020/2/19/21112570/univer... purports to list all of them, and doesn't list any failed studies.) The illusion of balance should not come at the expense of truth.
https://www.vox.com/future-perfect/2019/9/5/20849020/alaska-...
The fund is running out of money and there have been cuts to public services to keep it going because it’s now a political must have
https://alaskabeacon.com/2024/02/19/permanent-fund-corp-boar...
Your first link is just about politics. Per the data table on Wikipedia (https://en.wikipedia.org/wiki/Permanent_Fund_Dividend#Annual...), the massive increase described in that article didn't happen. In 2018, the PFD payout was $1600; in 2019, it was $1606. Not $6700.
If adequate, long-term funding can be acquired, I don't see why a UBI initiative should have these issues. The general principle "money represents other people's labour" applies, so if UBI prevents people from performing valuable labour within the economy (i.e., working for money), that could cause problems (though my instinct says that performing valuable services for others should be enough to avoid problems, regardless of whether payment is involved). Is there evidence of that happening in Alaska? doi:10.2139/ssrn.3118343 (also from "my" article) argues not.
For somewhere that isn't Alaska, the money would presumably have to come from taxes. I don't know enough to design that, but I don't see why it shouldn't be possible – especially if UBI has positive effects on the economy, as some theorists claim. (I don't tend to trust economic theorists.)
But since that’s hard, supporting the PFD, which is now a political must have, has required cutting public services which is what you said critics were unfounded in warning about.
If the knock-on economic effects of the PFD are putting more money into the system than it's taking out, it's obviously worth keeping (even when money is tight, unless that's expected to be temporary and certain other conditions are met). If the PFD is valued directly, then it's also worth keeping, regardless of the economic effects. We don't kill the elderly once they retire, even though that policy is of greater cost to the economy than benefit.
The naïve "it costs money that could be spent on other things" claim can be used to criticise any policy. The real question isn't whether it's good or bad, because we know there are circumstances that it's good. The question is, under what circumstances is UBI bad?
UBI seems to work in theory, so I'm mighty suspicious that it seems to work universally in practice. ("Economic theory just doesn't work that well," I continue to insist.) But the only failure mode I've seen, anywhere, is "it costs money" (e.g. https://www.employment-studies.co.uk/news/universal-basic-in...). All other criticisms appear hypothetical, so I don't get why we pay them so much heed.
The PFD is a political must have in the sense that voters are now telling the government they would rather cut other services to keep it
You previously disparaged critics for saying that there’s no evidence this would cut public services.
You are now changing your argument to say that it’s worth it in spite of that being shown to be a realized risk.
No they're not. Many countries have unelected leadership, and their political systems function. Ballots (and the rest of the democratic infrastructure) are something that people want, because they see them as good to have.
> The PFD is a political must have in the sense that voters are now telling the government they would rather cut other services to keep it
The cuts described in your first link didn't happen, because of public pressure against the cuts.
But let's pretend the Alaskan electorate did demand cuts elsewhere, to maintain or increase the PFD. If the budget reduces, "people would prefer to cut something else" is a sign of a service being valued by the populace. How is that an argument against the service?
Yes the did (and are doing). They didn’t get the 6.7k, but they did have to make cuts in large part because the payout was unsustainably high.
> How is that an argument against the service?
Don’t give a shit. I’m arguing against your claim that there’s no evidence UBI could lead to cuts in public services.
Unless $2k is enough to pay for a year of food, shelter and heating in Alaska.
Are there any that satisfy the conditions I consider essential, which I explained in my other comment?
Your criteria also omits the U
Your criteria also omits the U
Some of them would not either because they want to do nothing or because they want to invest in their skills; and so there would be a labor shortage and wages would rise and prices would rise and it would all equilibrium out to the point where people need to work to survive again (or UBI is raised to meet the B) and the whole economy gets a little less competitive with outsourced labor that doesn’t play this game.
That’s not a complete loss, mind you. A situation where poor people need to work to survive but a large percent of their income is guaranteed has merits.
Which of the listed studies actually test UBI (covering all basic needs, guaranteed to have no time limit)?
He was suggesting the people could be given "compute" which they then decide how to use. Though I don't deny this could be a loop-hole to create value in something like WorldCoin, my hope is that it would change the considerations of the people who have it.
When we think of UBI, we don't often think of money as the resource, but rather as the thing people use to buy things.
If that conversation changed to providing a resource which the person could use to earn, does that make the recipient more of a farmer or fisherman, rather than just being given a fish.
We rarely think of UBI as cash payment as something that the recipient would use as an allocation to create something for themselves.
Though I don't think this is a simple answer, I'd like us to think in these terms more, and Sam Altman has recently opened this window to my thinking.
It's easier to splurge with a card than it is with cash because it feels more abstract, particularly if it's credit.
Tieing UBI to working pretty much is counter to the word Universal.
Existing schemes which aren't UBI but are tied to working have caused a lot of problems where an individual is unable to work for some reason (possibly dealing with keeping access to the scheme) lose the benefits from the scheme and not cannot work as they needed those benefits.
Although I think the larger issue is that whatever resource you need to have enough of it and fungible enough to distribute. Distributing just one fish to 300 million Americans is actually nearly impossible. There's a reason the stimulus checks were done through the IRS; it's pretty hard to do things at this scale.
Money is actually no different really, except that we trade money directly goods. I look at this as a layer of abstraction which provides a re-framing. I don't think we're going to get rid of money, and will money move to be an allocation of AI credits? Maybe, but I hope not.
When someone is given $500/month, I believe the reaction for most people, outside of covid times when spending was a challenge, is, what will I spend this on. Given a resource that has an intermediate step and could be valuable to many people, the mentality may change to "what should I sell this for?".
The person can chose to donate it to a cause they feel valuable, sell it to the highest bidder, or multiple options in between.
That's my hope anyway, but if it was just dropped in tomorrow without the right social guardrails in place, I doubt it would work out like this.
As far as scalability, with hundreds of billions being spend on compute in order to train AI, I doubt scalability of distribution would be a major issue.
> They can use it themselves to earn, or they can be paid by others to use theirs. ...
> Given a resource that has an intermediate step and could be valuable to many people, the mentality may change to "what should I sell this for?".
Well, with your initial comment it sounded much more like you intended for people to get a resource (compute) that they themselves exploit to generate revenue (cash) as opposed to just giving people revenue (cash). Which to me defeats the whole point of UBI and is likely to be worse as whatever resource picked is in no way going to be exploitable by everybody or sellable for the same price as it was obtained by the government.
> As far as scalability, with hundreds of billions being spend on compute in order to train AI, I doubt scalability of distribution would be a major issue.
I guess with the covid stimulus being in the trillions, spending just billions wouldn't have the same scalability problems.
When you're not, you will either revolt or die. To think otherwise is rank naivety.
In a system with income-dependent benefits, poor people usually have the highest marginal tax rates. Often something like 70-90%, and sometimes above 100%. Not because of nominal taxes but due to losing benefits when their income increases.
Basic income can be largely cost-neutral. It would replace most common benefits, and it would be funded by most people paying a flat tax rate for all of their income. Typically you would see it only as an accounting trick: your nominal income tax is higher, but you receive a large automatic tax credit. People receiving benefits would have more incentives to work, which is the primary effect.
What you are suggesting is already done, and it's exactly what's wrong with the current system. When there are multiple separate low-income benefits, it's easy to accidentally create incentives against working. You try to work harder, but you lose 80% or 110% of the additional income to taxes and lost benefits. The latter because you cross one or more thresholds for the benefits.
Basic income avoids this with better coordination. There are no income thresholds for "low-income" benefits and no lower tax brackets for low to medium incomes. Everyone receives benefits in some technical sense, but most people pay them back through higher nominal taxes. In practice, most people would not receive basic income and would not pay any higher taxes. They would just do their tax calculations in a different way.
Never mind cutting off "those with jobs" would introduce all sorts of problems like we see with similar "benefit cliffs" for just basic welfare systems.
Or maybe things change when we consider hereditary wealth? With a below poverty line UBI, middle and upper classes could survive and maybe live quite comfortably with the income boost from investments, with the lower classes having to work to earn enough to pay their landlords.
I don't think we can rely on family or extended family any more. Even in countries with strong extended family and community support structures, people get left behind. And usually the people who need the most support, with mental illness or addiction or social problems driving others away.