Jeff Bezos’s Master Plan
theatlantic.com
theatlantic.com
There are however plenty of dystopian concepts in science fiction which his company seem to be propelling us toward:
* Compression of labour wages by replacing/augmenting with robots.
* Building an evil mega-corp (think Tyrell, Liandri, UAC, Weyland Yutani) that has its tentacles in every market.
* Facilitating the demise of human privacy laws
* Encouraging consumerism and subsequent pollution.
I am sure there are other themes in sci-fi that could be applied to Amazon.
Looking forward to my subdermal Alexa powered chip!
It's almost like a weird form of localized entropy reduction. Driving improvements in some areas whilst externalizing damage in others.
That said, I enjoy watching Expanse and look forward to s4 :)
Eliminating work by replacing it with robots is a social good, so long as the lost income is replaced somehow. Bezos is an outspoken proponent of UBI.
How can it be both?
I am sorry I dont follow.
I. Many "techies" are also science-fiction fans.
II. Space is where everything else is located.
> So much of earth vastly underpopulated and much easier to colonize.
For the next few minutes.
Either we stabilize population or we fill the Earth (exponential growth) or Nature stabilizes it for us (which is expected to be pretty gruesome if it comes to pass.)
If we want to keep growing we are going to have to do it in space. (Although that will only postpone the inevitable, albeit for perhaps millions or billions of generations.)
> What if an asteroid strikes earth?
That would suck, but it seems like a reason to colonize space.
> What if the space colonies all fail?
That would also suck. It is not a reason not to try though.
The odds are that all the colonies fail is real, just as the asteroid striking Earth, but the colonies on the long term perfectly hedge against a lot of failure modes.
[1]: https://www.youtube.com/channel/UCZFipeZtQM5CKUjx6grh54g
I hope he didn't actually say that, because it would betray a rather fundamental mistake in space travel - that of distance vs delta v.
His idea is simple: an asteroid is but a huge pile of resources, so instead of having to go e.g. to the belt and back (that's the major cost for space mining, especially from Earth to there — out of our gravity well), you essentially get a 'freebee' coming right to you. With enough tech, you can mine it and shrink it enough it's not even a hazard for earth, or deviate it enough to put it into orbit or slingshot it towards a handy location, whatever. It's rare enough an event (65 million years and counting?) that it would be a nice space-candy for a budding space civilization.
In such a slightly more advanced civilization, you'd see news like "Space corporations get ready as major asteroid coming towards Earth is expected to yield x tons of useful materials, equivalent to n billion dollars saved compared to belt mining." Not to mention asteroids coming from beyond our system could contain exotic materials scarcely found around here.
Basically, with a near earth asteroid you can make a Hohmann transfer there, mine the nice stuff, and a Hohmann transfer back home - and you can make sure to pick a platinum-iridium asteroid, or whatever it is that's most valuable. With an asteroid on a collision course, you probably won't have the years to spare to take such an efficient transfer, and you have to move the entire thing away from its collision course.
A valuable asteroid with a 10,000,000km-class intercept a decade out might be really, really convenient. But most potential impactors aren't that friendly.
> Not to mention asteroids coming from beyond our system could contain exotic materials scarcely found around here.
Well, rare-earth minerals are called that for a reason - they're rare on Earth. Platinum-iridium asteroids are nice! But that doesn't require coming from outside our solar system.
I genuinely had no idea the delta-v of so many asteroids was so small, compared to LEO. It fills me with 'educated hope', in regards to not-too-distant (in time) sustainable space industry — this and many other factors.
Regarding Isaac Arthur (I'd link the video, but there's so many), he definitely doesn't make that kind of rookie mistake, so I assume it's more of a tongue-in-cheek remark, to address popular concerns and show solutions — drive the point that a space-faring civilization is equipped to deal with such 'problems'.
About rare-Sun materials, I was more implying the possibility of finding isotopes or elements scarce in the belt and inner solar system — but you're right, even just the outer system is already full of useful stuff.
You've convinced me that it does seem pretty improbable that we'd get such a good near-earth candidate during the short window of time when it would really help us a lot, though, assuming we don't take another millennia to get out there (and even then).
Just curious, at the current pace (considering 1950-today) and barring any major stall/setback/boost (no WW3 etc), what would be your personal estimate for the first profitable mining operations? 10, 30, 100 years?
I guess I should say that there is one other advantage of an asteroid that's already on a collision course - if you can aerocapture instead of doing that propulsively, you only really have to pay the delta v to get to the asteroid in the first place. The problem is that the asteroid might have structural issues that intense acceleration exposes, and that people might be a tad leery of having a large asteroid on a course where if you miss by 100km you kill a billion.
> Just curious, at the current pace (considering 1950-today) and barring any major stall/setback/boost (no WW3 etc), what would be your personal estimate for the first profitable mining operations? 10, 30, 100 years?
I'll say "10 years after we get cheap ($500/kg class) launches" - which might be within the decade, or might be never. I think asteroid mining might still be profitable at far higher launch costs, but the risk and the startup costs would be far higher to the point that it probably won't happen.
Fair point! If unsure, we'd better not play this game too close from home.
> I'll say "10 years after we get cheap ($500/kg class) launches" - which might be within the decade, or might be never. I think asteroid mining might still be profitable at far higher launch costs, but the risk and the startup costs would be far higher to the point that it probably won't happen.
Thank you for sharing that perspective.
One game-changer is to build some sort of structure around or far enough from Earth's gravity (I'm partial to orbital rings); anything that can be built now using earth materials only (at least the basic stages), to kickstart a space industry. Though I suspect there are bigger political and economical challenges than technical ones on the way to cross this upward bound.
It's an appealing fantasy to which I was very strongly drawn in my youth, through O'Neill torii, T.A. Heppenheimer, von Braun, Clark, Asimov, and Heinlein.
I believe I've matured slightly.
Several other parts have grown in its place though.
Matured in favor of, or in opposition to, or in a different path altogether?
How do you plan to spread that maturity to others if what you're sharing is not clear? I grew up on the same diet of SF authors, and while I do acknowledge the need for work to be done here on earth, we shouldn't let that need tear down progress towards space. Especially when space expenditures are such a small part of what we as a country and the world spend our resources on.
The world, and the technology of it is in constant change; I really like your essay examples, in that the goalposts for your viewpoints are set quite thoroughly. If only everyone were so clear in their arguments...
Also a hedge against earth's collapse.
That's not true - we know far less about the oceans than we do about space.
>The expanding frontiers of space represent a fantasy - where there would be contant expansion of human possibility/potential.
That fantasy depends on the common science-fiction metaphor that space is just a bigger, wider ocean and that we just need to recreate the Age of Sail with space ships.
But barring the discovery of a warp drive, "constant expansion" into space is simply not feasible (see the Fermi Paradox.)
>Also a hedge against earth's collapse.
Well, that's already underway, so...
That said, re: Oceans -- I find it hard to believe that "pioneers" could build islands or sub-oceanic cities (Gungan-style), without a nation-state declaring their sovereignty over it. Maybe a floating flotilla of ships, like in Snow Crash? But, still, generally I think that nation-states have us just where they want us.
As Lovelock as said, currently the only answer is nuclear energy and that's largely rejected while the non-Western countries increase their fossil fuel use 'up to but not beyond' Western levels. They have a way to go and it seems not to involve cutting back whatever words emerge.
As much as I would like to go to space I don't think mass migration will be possible (unless we make some crazy technological breakthrough like anti-gravity or something.)
Being a multi-planterary species means that an existential threat on earth (I.e. asteroid, super bug, etc.) doesn’t wipe out all of humanity.
>Bezos’s enterprise upends long-held precepts about the fundamental nature of capitalism—especially an idea enshrined by the great Austrian economist Friedrich Hayek[, who] argued that no bureaucracy could ever match the miracle of markets, which spontaneously and efficiently aggregate the knowledge of a society.
>Amazon, however, has acquired the God’s-eye view of the economy that Hayek never imagined any single entity could hope to achieve. [...] With its logistics business—and its growing network of trucks and planes—it has an understanding of the flow of goods around the world. In other words, if Marxist revolutionaries ever seized power in the United States, they could nationalize Amazon and call it a day.
Perhaps we need to break up Amazon because communism? /s
Centralised planning is when a government planner decides the source, the destination and the good without giving final power to a consumer. Amazon doesn't have the final word on what people buy (or the destination of goods for that matter) so it isn't a centrally planned system in that sense.
Amazon is logistics, takeover would give the public a more extensive USPS plus digital infrastructure... it's not a revolutionary seizure of the means of production.
>it's not a revolutionary seizure of the means of production.
The revolutionary bit of Marxism was a political sideshow for Marx's anarchist supporters. The economic theory of Marx does not require a revolution, because according to that theory, the superior mode of economic production should eventually prevail.
That is why you have political movements like democratic socialism and social democracy. These are moderates who believe that Marx's theory of economic progression is correct, but who also believe in incremental change.
I think it's worthwhile to re-read the passage I quoted:
>Hayek[...] argued that no bureaucracy could ever match the miracle of markets, which spontaneously and efficiently aggregate the knowledge of a society.
This isn't just about physical logistics. Hayek is arguing that no centralised agent has perfect information. The author is arguing that Amazon is approaching the stage when it would have perfect information about the consumers it serves.
>Amazon doesn't have the final word on what people buy (or the destination of goods for that matter) so it isn't a centrally planned system in that sense.
Even if it doesn't have the final word, it does have quite a bit of say in the form of its recommender system.
Also, "communism" in the sense of Marx -- one of the two senses in which I was using the word -- is not necessarily a centrally planned system. Rather, it's an idealisation of what an economy would become if it achieved 100% automation, i.e. an economy operating at 100% efficiency.
You can see it in the structure of Marx's "historic materialism" argument: just as capitalism had displaced feudalism because it's more efficient, socialism would then displace capitalism, and communism would do the same for socialism. In Marx's hierarchy of efficiency:
Feudalism < capitalism < socialism < communism
So a communist economy, in Marx's sense, is also an economy in which there's perfect information. That can't possibly happen in the real world, of course, and Hayek correctly points this out about a centralised bureaucracy. But he's wrong about markets; no market is 100% efficient or operates in an environment of perfect information either.
This statement in the article is a huge misstatement.
First, Amazon only has information about the goods it sells. There are many things it doesn't sell, and many things that it can't sell. There are huge supply chains behind the products it sells that it does not see. There are also many things of value that are not exchanged with money.
Second, to the extent Amazon does have information about its customers, it has that information because of those customers' buying decisions. In other words, it has exactly the information that Hayek says a market participant has, and no more: the price at which other market participants will buy whatever quantity of one's products are sold.
>This statement in the article is a huge misstatement.
At least blame the right person: me. The above statement is a paraphrase of what I thought the author's argument was.
>First, Amazon only has information about the goods it sells.
>There are huge supply chains behind the products it sells that it does not see.
And it is in the process of acquiring more information. Amazon isn't just a retailer anymore: its main source of operating income is now AWS[0]. Also, dredmorbius has pointed out[1] that Amazon is doing vertical integration with its Amazon Business offering[2].
>There are many things it doesn't sell, and many things that it can't sell.
This is irrelevant, because these "many things" exist in a market outside of Amazon's "jurisdiction". Even in the historical centrally planned economies, these "many things" also existed.
>There are also many things of value that are not exchanged with money.
And they are getting rarer and rarer, because monetisation is now the Great Project.
>In other words, it has exactly the information that Hayek says a market participant has, and no more: the price at which other market participants will buy whatever quantity of one's products are sold.
Actually, Amazon has more information than a single market participant has. Since it is the market because it is a marketplace-as-a-service, it has exactly the information that Hayek says the sum total of all market participants have. This invalidates Hayek's argument that no centralised agent can possibly collect that much information. This point is what I believe the author of the article is trying to make.
[0] https://www.zdnet.com/article/in-2018-aws-delivered-most-of-...
Fair enough.
> Amazon has more information than a single market participant has. Since it is the market because it is a marketplace-as-a-service, it has exactly the information that Hayek says the sum total of all market participants have.
No, it doesn't. It has all the information that a large market participant has--i.e., a market participant that has a very large number of transactions with a very large number of other market participants. Hayek was perfectly well aware that such market participants existed (after all, such participants existed in his day--for example, Standard Oil).
But Amazon's information is still limited to information that a market participant can have: namely, as I said, the prices at which the various goods it sells are sold. It does not have any more information than that. It has information about a lot more transactions than most other market participants, but the information it has about each transaction is the same as what the other participant in the transaction has, no more.
And Hayek did mention monopolies or "coordinated industries" as a "halfway house" between centrally planned economies and free markets. Sure.
>But Amazon's information is still limited to information that a market participant can have: namely, as I said, the prices at which the various goods it sells are sold. It does not have any more information than that.
Actually, Amazon does collect and process information that a market participant wouldn't normally have: consumer preferences.
Hayek argued that a central planner wouldn't be able to suck up the decentralised information embodied in every market participant. However, what Amazon is doing is mining the data it has about a consumer's purchase history, as well as the responses the consumer gave to recommendations made by its recommender engine. That is a lot more information than what most market participants would normally have access to.
Now, Hayek has outlined what he thinks is needed to construct a "rational economic order" in the beginning of "The Use of Knowledge in Society":
1) "If we possess all the relevant information",
2) "if we can start out from a given system of preferences",
3) "if we command complete knowledge of available means".
Unlike the monopolies of old, which may have 1) to some extent, Amazon also has 2), and is trying to acquire 3) with its Amazon Business offering. So its capabilities are arguably an improvement over the capabilities of the old monopolies.
This is wrong. A market participant does normally have information about the preferences of other market participants. That's precisely the information conveyed by market transactions, buying and selling certain quantities of certain goods at certain prices. The only difference between Amazon and you or I is that Amazon engages in a lot more transactions, so it makes a lot more sense for Amazon to invest significant resources in collecting and analyzing its transaction data. But that doesn't make Amazon a successful central planner; it just makes Amazon a successful large market participant.
> Unlike the monopolies of old, which may have 1) to some extent, Amazon also has 2), and is trying to acquire 3) with its Amazon Business offering
Wrong again. Amazon has none of those three, nor has any other monopolist in history. Or indeed any central planner such as the government of the Soviet Union. Nobody has all relevant information. Nobody can take an entire system of preferences as given. And nobody has complete knowledge of all available means. All three of those requirements are extremely strong, and Hayek meant them to be that way. He didn't mean them as approximations; he meant them as literal exact requirements, which no real entity can possibly fulfill. Certainly Amazon doesn't.
> its capabilities are arguably an improvement over the capabilities of the old monopolies
I think this is true, but it's a much, much weaker claim than the one you have been making.
>I think this is true, but it's a much, much weaker claim than the one you have been making.
I think you need to stop identifying me with the author of the article. My original comment was to make fun of the author for his ludicrous comment about Marxists calling it a day after nationalizing Amazon, which got some Marxists riled up:
>Perhaps we need to break up Amazon because communism? /s
The comment you replied to in starting this conversation spelled out exactly the same argument you made above:
>So a communist economy, in Marx's sense, is also an economy in which there's perfect information. That can't possibly happen in the real world, of course, and Hayek correctly points this out about a centralised bureaucracy.
So the "much, much weaker claim" is the one I have always been making.
>>> Unlike the monopolies of old, which may have 1) to some extent, Amazon also has 2), and is trying to acquire 3) with its Amazon Business offering
>Wrong again. Amazon has none of those three, nor has any other monopolist in history.
It's exhausting to have to always qualify my statements. Of course I did mean "Amazon also has 2) to some extent, and is trying to make inroads into acquiring 3) to some extent".
>> Amazon does collect and process information that a market participant wouldn't normally have: consumer preferences.
>This is wrong. A market participant does normally have information about the preferences of other market participants. That's precisely the information conveyed by market transactions, buying and selling certain quantities of certain goods at certain prices.
No, this is wrong. Amazon collects information about the preferences of market participants through non-market non-priced transactions as well. Here's a non-exhaustive list of such information collection:
- information from consumer responses to recommendations made by the recommender system
- information from consumers immediately after they've clicked on a link supplied by an Amazon associate, which then directs them to a product page.[0] Yes, there is a market transaction between Amazon and its associates, but at the time of the click, the information isn't generated from a market transaction between Amazon and the consumer.
- information supplied by consumers themselves through Amazon Lists, listing explicitly their preferences for goods and services.[1]
- information supplied by consumers themselves about their social network and preferences through Amazon Wish List.[1]
This is evidence for suggesting that "Amazon also has 2) [to some extent]", possibly to a greater extent than any monopolies of the past century has ever had.
To the extent this makes a difference, because of Amazon's efforts to monopolize various market segments by undercutting competition, it decreases the amount of information Amazon has about other market participants. A perfect monopolist, by setting monopoly prices, decreases the value of the price information it gets through transactions, because there are no other competitors offering the same goods at different prices to compare to, so the prices of the monopolist's transactions convey less information about the preferences of other market participants. That makes the market less efficient.
>To the extent this makes a difference
It does, because you're arguing as if Amazon sells stuff. It doesn't (just) do that: it's a market-maker that provides a venue for other vendors to sell stuff.
So it's already an error to talk about Amazon as a monopolist, when Amazon isn't the only marketplace out there with a global reach. But you're also making a category error, because you wrote:
>A perfect monopolist, by setting monopoly prices, decreases the value of the price information it gets through transactions, because there are no other competitors offering the same goods at different prices to compare to, so the prices of the monopolist's transactions convey less information about the preferences of other market participants.
Amazon doesn't set prices per se, at least not for vendors in its marketplace that are not itself. So it doesn't do whatever you think it's doing when you were writing that paragraph. Certainly there are "other competitors offering the same goods at different prices to compare" with, because Amazon's marketplace include other vendors. But Amazon is running a marketplace, not (just) being a vendor. There's a difference in category there.
That's not what a market maker is. A market maker is a market participant that guarantees to buy or sell something if no other buyer or seller can be found.
Providing a venue for others to sell stuff is a straightforward market transaction: Amazon sells its services to others for a price. There's nothing special about it, and standard Hayekian economics accounts for it just fine.
> Amazon doesn't set prices per se
To the extent this is true (as you say, for vendors that are not itself), it just reinforces the point I made above: that what Amazon is doing as a "marketplace as a service" is just a straightforward market transaction, selling services for a price, and does not give Amazon any sort of special status.
> Amazon is running a marketplace, not (just) being a vendor.
Running a marketplace, if it does not mean being a monopolist (which you are insisting it does not), is being a vendor, a vendor of particular services.
>That's not what a market maker is. A market maker is a market participant that guarantees to buy or sell something if no other buyer or seller can be found.
Right, but the similarity between a market maker in finance, and a marketplace maker in ecommerce, is that both of them have a lot more information than an ordinary buyer or seller. It doesn't matter if Amazon doesn't really act as a counterparty: the point is about how much knowledge it has, compared to the average participant.
>Running a marketplace, if it does not mean being a monopolist (which you are insisting it does not), is being a vendor, a vendor of particular services.
I was pointing out that there are two levels we're looking at:
1. Amazon as a provider of a marketplace
2. Amazon as that marketplace and what it entails about how much knowledge it possesses.
Recall that this is in the context of Hayek's arguments in The Use of Knowledge in Society, because the article's author referenced that explicitly.
Now, when I said Amazon isn't a monopolist, this is in terms of 1. Amazon is definitely not the only provider of an ecommerce marketplace with a global reach. In that sense, it's not a monopoly.
But Amazon is a "monopoly" in the sense that Hayek alluded to in his essay, i.e. in terms of 2. Hayek wrote in that essay:
The halfway house between the two, about which many people talk but which few like when they see it, is the delegation of planning to organized industries, or, in other words, monopoly.
Technically, it is still an oligopolist, since it is not the sole provider of ecommerce in America, but "organized industries" sounds like an apt way to describe what Amazon is doing. And it's in this particular sense of Hayek's that I referred to Amazon as a "monopoly".
Hayek's basic premise was that national-scale central management was impossible. Amazon and other megacorps put paid to this.
Also - they deal in finished (consumer) goods. There is a whole lot of coordination before that stuff gets assembled and ready for Amazon or one or their marketplace suppliers to ship it.
Maybe they've integrated the demand chain logistics, but not the integration supply chain. And they don't dictate that consumers get to purchase one and only one brand of whatever product (like peanut butter).
(Although I knew about it and stand by my assertion, because...)
I still can't conceive of using that to build a (say) iPhone.
what if it was an ai instead?
Marxists could not call it a day after nationalizing Amazon because the people who make the goods that Amazon sells would still be alienated from the goods they produce, and Amazon's workers would still be alienated from the products of their own labor.
I don't think the author meant that reference to Marxism as anything more than a rhetorical flourish. His target, in that context, was Hayek's argument from incredulity against central planning in The Use of Knowledge in Society.
>Marxists could not call it a day after nationalizing Amazon because the people who make the goods that Amazon sells would still be alienated from the goods they produce, and Amazon's workers would still be alienated from the products of their own labor.
As I understand it, this alienation -- as you've described it here -- seems to have its roots in the division of labour in society, which seems to me to pre-date industrial capitalism. If that's so, then the work of Marxists can never be done, because work itself would have to be totally eliminated for the division of labour to vanish.
Isn't that an argument for Amazon is de-facto a monopoly at this stage?
Because there is evidence that Marxists had tried, and failed, to do what Amazon has now succeeded. Project Cybersyn comes to mind:
https://en.wikipedia.org/wiki/Project_Cybersyn
https://www.newstatesman.com/world/2018/08/project-cybersyn-...
However, what Amazon has done is just characteristic of "organized industries", which Hayek conceded in his essay, The Use of Knowledge in Society, as being a "halfway house" between Soviet-style planning and free markets.
Space colonization is a form of greed: it's basically saying this world is not enough. So I don't see this as a rationalization: in fact, the article heavily suggests that his dream of space colonization was what led to his greed.
If an extinction-level event occurs on earth, humanity can continue to flourish and populate the cosmos rather than go the way of the dinosaur.
"Greed" and humanitarianism are not mutually exclusive. In fact, they can be mutually compatible, depending on what your definition of "greed" is.
From an economic point of view, humanitarianism can serve the economic "greed" of the rich that's often decried. This is because both of these can be rooted in the realisation that humans have intrinsic economic value, so that the loss of a human life represents a loss of a lifetime of economic value to society.
In the case of Bezos, it's quite clear from the article that his humanitarianism is intrinsically linked to his "greed":
>When Bezos describes these colonies—and presents artists’ renderings of them—he sounds almost rapturous. “This is Maui on its best day, all year long. No rain, no storms, no earthquakes.” Since the colonies would allow the human population to grow without any earthly constraints, the species would flourish like never before: “We can have a trillion humans in the solar system, which means we’d have a thousand Mozarts and a thousand Einsteins. This would be an incredible civilization.”
"We can have a trillion humans [...] This would be an incredible civilization." Sounds familiar? This is essentially the siren song of the Chinese market, magnified a thousandfold.
Why do you think that of him?
And by the way, was that happening while you were working at Amazon?
I have been working for Amazon for over 2 years and I dont live in poverty or misery. In fact, thanks to my pay at Amazon (2x the minimum wage in my area) I can live quite comfortably. Thanks to my job I dont stress out when I think about my student loans. And thanks to the innovation which was implemented with robots, my job is a lot easier that what it would be without the assistance of the electronics we use.
- Miguel FC Ambassador
Only half sarcastic...