Although not mentioned in either the article, or the essay it links to [0], Sam Altman is also a big proponent of Universal Basic Income[1]. My understanding is that part of his rationale for UBI is that certain types of jobs will disappear, partially due to AI.
Fyi Bitcoin can be manipulated and your wealth effectively redistributed if more than 51% of miner nodes act adversarially. (By some calculations it's only 1/3 required). If entities with billions want to hack bitcoin they theoretically could.
Surprising many don't know that Bitcoin has these takeover risks.
Not saying Bitcoin is worse just that its vulnerable... Since we don't know for sure who controls what hash power, it's even possible that quiet manipulation already takes place.
On brighter side blockchain projects that require memory bound hashing algorithms are less vulnerable to takeovers since massive ASIC farms don't work afaik. Ethereum is one such approach (but ofc when they decided to undo transaction history after TheDao hack, it isn't immutable either).
I disagree with the "only", yes if a capital-intensive takeover happened they cannot take your bitcoins from your wallet, but they can still straight up destroy bitcoin or quietly enrich select parties through double spends as btc inflates.
Fwiw the history can be re-written if over a long time a miner produces a longer chain due to hash power advantages in secret; when released it would by default invalidate all transactions since the secret mining began.
Those entities with billions COULD do that, fully aware that it would utterly wipe out their billions in BTC assets. So why in the world WOULD they choose to do so?
When it comes to having control over a society I don't think some of these powerful groups care about wiping out a few billion. See covid-19 extended economy shut downs.
> The beauty of Bitcoin is that my wealth will never be redistributed
I'm skeptical. That's presuming Bitcoin stays stagnant, but that stagnation could also mean its downfall.
Also "wealth" locked up in Bitcoin is only as good as your onramps and offramps into the mainstream economy, and it's quite possible that those can be legally closed off, putting you in the position of someone like a drug dealer if you want to access it.
I think he shouldn't try for a national, political format. Instead, I think he should form a sort of member-owned corporation. Make the AI, make the money, pay dividends.
> “The Green Leopard Plague” is the story for which this collection is named. It is the best work I’ve seen in explaining how economics and taxes would work in a post-scarcity world. How do you make people pay taxes and do dirty work when they don’t have to work to eat? And the narrative lacing around the academic discussions is a quest to end world hunger by genetic engineering, the namesake plague, and a failed youthful romance in the post-singularity future.
The Golden Age by John C. Wright is the start of a trilogy that deep dives into a post-scarcity solar system. It doesn't directly analyze the economy, but a big theme is what happens when someone is outcast from society, and has to start over with his bare hands and feet, scrounge up his own tech and off-grid wealth, and negotiate deals and arrangements with exotic beings on Earth and elsewhere.
Sam's idea of taxing market cap would cause much more than the 15% depreciation than he seems to expect. Any company with a P/E of 40 or more immediately would be losing money. Companies with P/E of 20 would lose half their profit. That's most companies. I would expect a 50% drop of the S&P if this policy was enacted. Even if this tax is somehow enacted it would only collect half as much as Sam expects.
Companies could do this by creating new shares, not having to pay in cash. This could be automated by the government, each year all companies listed on a USA exchange create 2.5% new shares which are then given to the government. The government then gives these shares evenly to all adults in the USA. It's then up those people if they want to hold them or sell their fraction on existing marketplaces (which I assume in 10 years will mostly support fractional shares, especially if this is coming).
There is effectively no difference between a dividend and a buyback. You're trying to claim that doing a reverse buyback is different than just paying taxes, but they would have the same effect.
UBI is here (perhaps temporarily), and unevenly distributed in terms of fiscal and monetary allocation.
AI is here, and in the hands of a few tax-evasive global corporations. So far AI has been good for: addictive attention farming, polarization, and social control.
I'm a fellow techno-optimist, but the solutions Sam brings up don't feel effective, or at least sufficient. Our direction is ugly, and Sam isn't calling out the big players to change their course.
Perhaps we need Universal Basic Compute: any corporation beyond a certain compute capacity (in FLOP/s) must allocate 50% to public utility. From there, individuals can vote to determine how they algorithmically represent themselves.
> By 2000, the machines will be producing so much that everyone in the U.S. will, in effect, be independently wealthy. With Government benefits, even nonworking families will have, by one estimate, an annual income of $30,000-$40,000 (in 1966 dollars). How to use leisure meaningfully will be a major problem...
First paragraph of that article is incredibly spot-on! The closing paragraph was also quite prescient, no matter what side of the political parties you root for:
> There are some who gloomily expect a society run by a small elected elite, presiding over a mindless multitude kept happy by drugs and circuses, much as in Huxley's Brave New World
Didn’t BillG suggest the labor of robots and other artificial beings be taxed? In a way, we might not have any other choice if we want to avoid total societal collapse or communism.
I doubt that people behind AI-based means of production will be altruistic enough to share the wealth created by AI systems. We are not sharing the wealth created by trading bots, or other forms of automation. If anything, it has helped worsen the concentration of wealth problem.
My prediction is that automation will create a neo-feudal society that will slowly transition into a humanless society. Wealth created by automation will be used to buy land and automated defense systems, and all the wealth, territory and power will concentrate in few hands.
Automated defense systems will not have morality, will not make any questions and will just do what they're asked. Their allegiance will be to the people buying them, and that allegiance will be total and unconditional.
Ultimately, all the wealth asymmetry will be converted into a territorial and military power asymmetry, and once this conversion is completed, the elites will enslave humanity and place limits on their reproduction rate. Why? because they won't want to share the natural resources with random people.
For now, the name of the game is to accumulate as much money as possible until that conversion is possible and invest in the advancement of AI.
Then, there will be a competition between the remaining people, that will force them to develop more powerful AI to prevail over each other. After certain point, more decisions will have to be delegated to the AIs in order to prevail making the AIs impossible to control. This is when AI augmented humans will become human augmented AIs. The influence of AIs will grow and then AIs will mutate humans to suit their needs, doing to us the same we do to cattle: artificially selecting individuals favoring the traits they like, turning us into the equivalent of farm animals.
In a post scarcity society, it doesn't make so much sense to talk of dollars.
If these heady futures of AI do come to fruition, then the cost of "stuff" trends towards zero. We've sort of seen this already playing out since the industrial revolution, and more recently outsourcing manufacturing to countries with very cheap labour. It used to be that it would take a week's wages to buy a pair of pants. Now you can buy a cheap pair for an hour of minimum wage.
So what has happened? People's expectations have increased, and more and more of people's incomes get shifted to focus on the physically scarce e.g. desirable land.
Stuff (and most labour value) will trend toward zero, and land will trend toward the infinite.
If we do get to a "post scarcity" economy, we really need to ensure no one person/company controls too much land. At a certain point, most labour will not be exchangeable for enough $$ to trade for land, and you'll have very little economic mobility for the majority.
At this point one might expect revolution, but AI would also upset the balance of physical projection of power the masses currently have over the few.
Former McDonald’s CFO, Harry J. Sonneborn, is even quoted as saying, “we are not technically in the food business. We are in the real estate business. The only reason we sell fifteen-cent hamburgers is because they are the greatest producer of revenue, from which our tenants can pay us our rent.”
If a person receiving UBI (which Sam Altman is a big supporter of), publicly espoused views such as neo-Nazi views, would anyone say that this person should continue to receive UBI? What if they lived in an inexpensive area and the UBI was sufficient to allow full-time focus on such views?
I've heard sincere leftist advocates for universal healthcare and universal basic income argue that it should be truly universal and they said yes, neo-nazis should get it too. Means-testing and adding criteria is starting to become an anathema concept to those who are ideologically committed to universal state benefits. However, that kind of consistent integrity may not extend to the politicians implementing it or the wider public.
> would anyone say that this person should continue to receive UBI?
I would.
There are two ways to look at this:
1. We are supporting this person through UBI. We don't want to support neo-Nazi people. Therefore they should not receive UBI.
2. Neo-nazism is an issue we want to solve. Neo-nazism only dies when no significant portion of society shares that view anymore. Notice that this is very different from "when a majority of society thinks it should be eradicated" (which is where we are now).
If you focus in the latter view, cutting UBI to publicly known supporters won't help because:
- It would foster the "us vs them" holy war between supporters and society at large. This will only strengthen that minority.
- It poses significant (i.e.: expensive to solve) logistical issues. How do you detect neo-nazi supporters? Who makes the decision? How can such decision be appealed?
I see neo-nazism as a virus and education as a vaccine. Cutting UBI to identified supporters is akin to cutting life support to people displaying advanced symptoms. Would that really help?
This is just ridiculous. Unless AI robots are physically wiping our asses, life quality is still a logistical issue and not a money issue.
I can give every citizen 1 million dollars and theres still only so many homes, so many nurses, so much food, so many education spots, etc. And things still need to be repaired, cleaned, cared for, etc. There is never going to be a period where the govt pays people and everyone sits on their asses because things do not get produced and therefore humans do not receive. Can AI help efficiency...sure, but its not going to provide everyone with a dream life of doing nothing and prospering.
The only way life quality for humans goes up is though greater production output (which AI can help). Giving people money does essentially nothing except cause inflation for resources as they are fixed in quantity.
65 comments
[ 3.2 ms ] story [ 127 ms ] thread[0] https://moores.samaltman.com/
[1] https://ycombinator.wpengine.com/basic-income/
The comment I replied to indicated violence being used to redistribute wealth. Is one not allowed to defend oneself against violent mobs?
Surprising many don't know that Bitcoin has these takeover risks.
I suspect this line of thinking will usually end with "world economies are backed by military groups.", to which I have no rebuttal.
On brighter side blockchain projects that require memory bound hashing algorithms are less vulnerable to takeovers since massive ASIC farms don't work afaik. Ethereum is one such approach (but ofc when they decided to undo transaction history after TheDao hack, it isn't immutable either).
Re-writing historical transactions from before the 51% attack is impossible, unless he can somehow re-write the hard drives in each node's computer.
https://en.bitcoin.it/wiki/Weaknesses#Attacker_has_a_lot_of_...
Fwiw the history can be re-written if over a long time a miner produces a longer chain due to hash power advantages in secret; when released it would by default invalidate all transactions since the secret mining began.
I'm skeptical. That's presuming Bitcoin stays stagnant, but that stagnation could also mean its downfall.
Also "wealth" locked up in Bitcoin is only as good as your onramps and offramps into the mainstream economy, and it's quite possible that those can be legally closed off, putting you in the position of someone like a drug dealer if you want to access it.
Right now, AI is mostly "fugazzi".
> “The Green Leopard Plague” is the story for which this collection is named. It is the best work I’ve seen in explaining how economics and taxes would work in a post-scarcity world. How do you make people pay taxes and do dirty work when they don’t have to work to eat? And the narrative lacing around the academic discussions is a quest to end world hunger by genetic engineering, the namesake plague, and a failed youthful romance in the post-singularity future.
https://www.goodreads.com/book/show/207410.The_Golden_Age
https://www.inkshares.com/books/trekonomics
Sounds ideal for curling up to read with a cup of "Tea, Earl Grey, Hot" from the replicator.
> What follows is an idea in the spirit of a conversation starter.
AI is here, and in the hands of a few tax-evasive global corporations. So far AI has been good for: addictive attention farming, polarization, and social control.
I'm a fellow techno-optimist, but the solutions Sam brings up don't feel effective, or at least sufficient. Our direction is ugly, and Sam isn't calling out the big players to change their course.
Perhaps we need Universal Basic Compute: any corporation beyond a certain compute capacity (in FLOP/s) must allocate 50% to public utility. From there, individuals can vote to determine how they algorithmically represent themselves.
> By 2000, the machines will be producing so much that everyone in the U.S. will, in effect, be independently wealthy. With Government benefits, even nonworking families will have, by one estimate, an annual income of $30,000-$40,000 (in 1966 dollars). How to use leisure meaningfully will be a major problem...
> There are some who gloomily expect a society run by a small elected elite, presiding over a mindless multitude kept happy by drugs and circuses, much as in Huxley's Brave New World
Not just that, there is a limit to bucket-sorting, which is the only thing current AI is good for.
And you think they are going to redistribute profits, when they don't even pay taxes?
I have an NFT to sell you then.
My prediction is that automation will create a neo-feudal society that will slowly transition into a humanless society. Wealth created by automation will be used to buy land and automated defense systems, and all the wealth, territory and power will concentrate in few hands.
Automated defense systems will not have morality, will not make any questions and will just do what they're asked. Their allegiance will be to the people buying them, and that allegiance will be total and unconditional.
Ultimately, all the wealth asymmetry will be converted into a territorial and military power asymmetry, and once this conversion is completed, the elites will enslave humanity and place limits on their reproduction rate. Why? because they won't want to share the natural resources with random people.
For now, the name of the game is to accumulate as much money as possible until that conversion is possible and invest in the advancement of AI.
Then, there will be a competition between the remaining people, that will force them to develop more powerful AI to prevail over each other. After certain point, more decisions will have to be delegated to the AIs in order to prevail making the AIs impossible to control. This is when AI augmented humans will become human augmented AIs. The influence of AIs will grow and then AIs will mutate humans to suit their needs, doing to us the same we do to cattle: artificially selecting individuals favoring the traits they like, turning us into the equivalent of farm animals.
If these heady futures of AI do come to fruition, then the cost of "stuff" trends towards zero. We've sort of seen this already playing out since the industrial revolution, and more recently outsourcing manufacturing to countries with very cheap labour. It used to be that it would take a week's wages to buy a pair of pants. Now you can buy a cheap pair for an hour of minimum wage.
So what has happened? People's expectations have increased, and more and more of people's incomes get shifted to focus on the physically scarce e.g. desirable land.
Stuff (and most labour value) will trend toward zero, and land will trend toward the infinite.
If we do get to a "post scarcity" economy, we really need to ensure no one person/company controls too much land. At a certain point, most labour will not be exchangeable for enough $$ to trade for land, and you'll have very little economic mobility for the majority.
At this point one might expect revolution, but AI would also upset the balance of physical projection of power the masses currently have over the few.
Interesting to think about!
We're already in a post-scarcity society for:
.. and probably other sectors? Each has their own scarcity dynamics and network effects, but the price of "stuff" has indeed trended towards zero.US is currently experiencing a housing/land shortage in desirable regions outside of big cities.
https://blog.wallstreetsurvivor.com/2015/10/08/mcdonalds-bey...
Defining a list of topics and views that would take your UBI away is the ugliest big brother/dystopian future I can imagine. Who manages the list?
I would.
There are two ways to look at this:
1. We are supporting this person through UBI. We don't want to support neo-Nazi people. Therefore they should not receive UBI.
2. Neo-nazism is an issue we want to solve. Neo-nazism only dies when no significant portion of society shares that view anymore. Notice that this is very different from "when a majority of society thinks it should be eradicated" (which is where we are now).
If you focus in the latter view, cutting UBI to publicly known supporters won't help because:
- It would foster the "us vs them" holy war between supporters and society at large. This will only strengthen that minority.
- It poses significant (i.e.: expensive to solve) logistical issues. How do you detect neo-nazi supporters? Who makes the decision? How can such decision be appealed?
I see neo-nazism as a virus and education as a vaccine. Cutting UBI to identified supporters is akin to cutting life support to people displaying advanced symptoms. Would that really help?
I can give every citizen 1 million dollars and theres still only so many homes, so many nurses, so much food, so many education spots, etc. And things still need to be repaired, cleaned, cared for, etc. There is never going to be a period where the govt pays people and everyone sits on their asses because things do not get produced and therefore humans do not receive. Can AI help efficiency...sure, but its not going to provide everyone with a dream life of doing nothing and prospering.
The only way life quality for humans goes up is though greater production output (which AI can help). Giving people money does essentially nothing except cause inflation for resources as they are fixed in quantity.
2. $13.5k/year still keeps you below the poverty line.
In other words, if there is no job to be had and your $13.5k "AI grant" is your sole income, you're still fucked.
The previous comments about "scraps" seem to be pretty spot on.