Can AI Solve the Socialist Calculation Problem?

Summary

---

The Human Action Podcast

Episode: Can AI Solve the Socialist Calculation Problem?
Host: Dr. Bob Murphy (Mises Institute)
Date: March 27, 2026

---

Overview

In this episode, Dr. Bob Murphy explores whether advances in artificial intelligence (AI) can overcome the classic "socialist calculation problem" debated by Ludwig von Mises and Friedrich Hayek. Murphy dissects the original arguments from the Austrian tradition, carefully distinguishing between Mises's calculation problem and Hayek's knowledge problem, and evaluates if AI genuinely addresses these critiques of central planning.

---

Main Discussion Points

1. Distinction Between Mises’s Calculation Problem & Hayek’s Knowledge Problem

• Murphy opens by clarifying a common confusion—even among Austrian economists—about the difference between Mises’s and Hayek’s critiques of socialism ([00:13–06:55]).

• Mises's Calculation Problem:

  • Not merely about incentives or corruption, but a fundamental inability of socialism to rationally calculate resource allocation due to lack of genuine market prices for means of production ([05:31]).
  • “What [Mises] put his finger on was a completely independent problem … that the planners, in order even ex post to say, did we come up with an efficient use of society's scarce resources...nobody would know ... whether that constituted inefficient use ... because they don't have the costs, right?” —Dr. Bob Murphy ([19:19])

• Hayek's Knowledge Problem:

  • Focuses on the issue that knowledge relevant to economic decision-making is dispersed and often tacit, making central transmission and processing impractical ([06:35–09:34]).
  • "Hayek argued that the price system, in a very economical fashion, communicates the relevant signals to all the people who have to make decisions." —Dr. Bob Murphy ([15:35])

2. Common Problems with Socialism (Incentives, Corruption, Power)

• Murphy distinguishes Mises's calculation critique from other well-known socialist pitfalls:
  • Incentive Problem: Equal pay ignores differing effort and unpleasantness of jobs, undermining motivation ([01:56–03:08]).
  • Corruption/Concentration of Power: Centralization enables abuse and restricts dissent ([03:08–04:39]).

3. Roots of the Economic Calculation Debate

• Historical Contours:
  • Mises’s 1920 article outlined the impossibility of economic calculation without market prices ([17:01–21:20]).
  • Hayek’s later contributions (e.g., "The Use of Knowledge in Society" [1945]) reframed the issue in terms of distributed knowledge and practical infeasibility ([09:03–10:13], [22:00–24:00]).
• Socialist responses (Dickinson, Lange) tried to show that calculation could be done via mathematical methods and trial/error adjustments to price signals ([24:00–30:00]).

4. Can AI Solve the Socialist Calculation Problem?

• The discussion pivots to a recent (2023) Twitter thread by economist Daron Acemoglu, which argues that AI may overcome the knowledge aggregation barrier in Hayek’s critique, thereby improving central planning ([06:55–12:13]).

• Murphy concedes AI could help forecast consumer demand and aggregate information, but stresses:

  • Even if AI perfectly collected and processed all technical and preference data, Mises’s core challenge remains unaddressed ([18:08–19:19]): socialism lacks genuine market prices, making rational resource allocation and post-hoc evaluation of efficiency impossible ([19:19–20:10]).
  • The “appraisement” entrepreneurs provide is not just calculation or knowledge, but a unique process only possible in the institutional context of market prices and decentralized property ([33:31–36:10]).

• Memorable Analogy ([36:10]):

  • “When we use a thermometer to measure the temperature inside an oven ... There's no doubt that the oven really has a temperature regardless of our measuring it ... In contrast, if we ask, how much economic value does this oven possess, then that's a fundamentally different question.” —Dr. Bob Murphy

5. Critique of the AI Solution

• Murphy points out ([44:00–49:00]):

  • At most, AI allows central planners to simulate what a decentralized market would do; it does not substitute for the market process or create real market prices.
  • All the resources devoted to AI simulations are themselves subject to opportunity costs and would be better allocated in a genuine market system.
  • Advances in AI are only as valuable as they are embedded in real market feedback—profit and loss signals—which guide their development and use ([48:25]).

• Core Quote ([49:00]):

  • "All these advances that companies are making right now in the real world with AI, those improvements are only able to be steered properly … if their owners are embedded in a market economy and are responding to profit and loss signals." —Dr. Bob Murphy

6. Final Remarks and Further Resources

• Murphy references his written works and Joe Salerno’s article for a deeper dive into the theoretical distinctions ([37:40–39:00]).
• He briefly mentions his own article using Cantor’s Diagonal Argument to show that—on a mathematical level—the calculation problem may be insurmountable even in principle ([50:30]).

---

Notable Quotes & Moments (with Timestamps)

• Calculation v. Knowledge Problems:

  • “That is not the same thing as Mises calculation argument ... the problem is nobody would know … whether that constituted inefficient use ... because they don’t have the costs, right? They don’t know the value of the inputs that were used up compared to the value of the outputs that were produced. That’s the fundamental problem.” —[19:19]

• Hayek’s Summary of the Knowledge Problem:

  • “The people don’t need to know, why is copper more scarce right now? They just need to know that it is. And so Hayek argued that the price system, in a very economical fashion, communicates the relevant signals to everyone who has to make decisions.” —[15:35–15:40]

• AI as Simulation, Not Solution:

  • "The only way that the AI systems could approximate the problem is they would have to simulate ... a market economy on this, you know, alt earth, this alt capitalist earth..." —[44:00]

• Market Prices as Unique Information:

  • “The market price is the information. It's not that it's just, you know, a surrogate or a proxy for something else.” —[39:15]

• Final Rebuke of Technological Optimism:

  • "Suppose that's all true. That doesn't prove, oh, so actually socialism could get by and that ... modern technology has rendered their objections obsolete. No, because still the AI helps Amazon ... embedded in a capitalist framework. ... You can't send the guy in Tallahassee more stuff than he's able to afford. That wouldn't be efficient. That would be a waste of resources." —[48:25]

---

Segment Timestamps

• Distinctions: Mises vs. Hayek — [00:13–06:55]
• Three Classic Socialist Problems — [01:56–05:28]
• Historical Debate & Key Figures — [17:01–30:00]
• How AI “Solves” (or Not) the Problem — [06:55–19:19], [44:00–50:00]
• Economic Calculation Analogy — [36:10]
• AI, Opportunity Cost, and Profits — [48:25]
• References to Further Reading — [37:40–39:00], [50:30]
• Closing Thoughts — [50:30–52:03]

---

Style and Closing Thoughts

• Dr. Murphy maintains an accessible yet rigorous tone, referencing both economic theory and real-world technology (Amazon, drones, AI). He is careful to avoid straw-manning opponents, reiterating several times that both Mises’s and Hayek’s critiques have merit, but represent fundamentally different challenges to socialism.
• The episode is rich with analogies and historical context, making complex ideas graspable even for non-specialists.

---

For further exploration:

• Joe Salerno’s 1994 article on appraisement
• Murphy’s own articles on Mises.org and the Quarterly Journal (“Cantor’s Diagonal Argument”)

---

In summary:
AI may improve information gathering and simulate some market dynamics, but cannot solve the foundational calculation problem identified by Mises, nor replace the crucial role of genuine market prices in coordinating an efficient economy. The distinction between “knowledge” and “calculation” critiques is foundational yet often overlooked; this episode is a detailed yet digestible guide to understanding why the Austrian critique of socialism persists—even in the AI age.

Transcript

[A]

This is the Human Action Podcast where we debunk the economic, political and even cultural myths of the days. Here's your host, Dr. Bob Murphy.

[B]

Hey everybody. Welcome back to the Human Action Podcast. This episode I am going to do a deep dive, at least for podcast intensity levels, into the socialist calculation debate. And specifically I want to flesh out, among other things I'm going to try to accomplish here is the distinction between the Mrs. And calculation problem and the Hayekian knowledge problem. And I think some people even card carrying Austrian economists conflate those two things and they'll say, oh, you know, it's Mises and Hayek were basically arguing the same thing. Maybe they'll say Hayek was just kind of fleshing out some of the details of Mises original salvo against the socialists. And I don't think that's correct. All right, and just part of the context here, just in case you're not aware, you're innocent lamb to these debates, there are different factions in the Austrian community and some people really like Mises a lot and some people like Hayek a lot. And depending on who your hero is, you might, you know, hold up one of them on a pedestal and say the other one is overrated. And okay, so that, that plays into some of this stuff. But regardless of how some partisans might take my claims and run with them, I'm going to say there's pretty demonstrably a difference between what Mises laid out and what we call the calculation problem with socialism versus what Hayek spelled out and what we could call the knowledge problem. Okay, they're both valid thing to me it's, it's a broader attack, right. Just like there's an incentive problem with socialism, right.

[C]

That geez, if everyone gets paid according to formula, like you know, how many kids do you have? Or are you a good person? Something like that.

[B]

Not how hard, how much work did you do and or, or how distasteful

[C]

is it, you know, like if garbage men get paid the same as librarians, you know, that kind of thing, then maybe you're going to have a problem getting people to go empty the really disgusting dumpsters and things like that. Right? Whereas in a free market, yeah. If there's some jobs that are dirty, then they earn more and that's how they motivate people to go do that stuff.

[B]

Right.

[C]

Or to run a certain company. Geez, you just geez, if you want to run that restaurant, do it right. You basically have no personal life for the next three years until that thing's

[B]

really up and running and you got

[C]

your clientele, and then maybe you can dial back. Well, if you get paid the same, whether you do that or you just go and watch kids, you know, at the daycare center and play with blocks with them or whatever, and you just do that from nine to five and you go home and have your weekends to yourself. Don't expect people to put in the effort, the grueling effort required to get that restaurant up and running if they don't get paid accordingly.

[B]

Right?

[C]

And we could call that the incentive problem, but that's not what was in Mises critique of socialism. Okay? And so by pointing that stuff. Oh, sorry, let me just mention one last thing. A completely independent problem for all this stuff, you might call the corruption problem with socialism or the power, you know, the concentration of power or the, you know, genocide problem, that then this is borne out by empirical reality that to have a centrally planned economy, among other issues, that places an incredible amount of power in the hands of a relatively few people. And so power corrupts. You know, this is pretty standard stuff here in terms of human behavior. And so among other issues, you could say no outright central planning. Old school socialism is a bad idea. We should not go down that path. If for no other reason than if the wrong people seize control, they can just slaughter millions of their political opponents, just, you know, or assigning them to, quote, jobs in Siberia, you know, to use a cliche.

[B]

Right.

[C]

And, oh, if I'm in charge of what schools people go to and how much food they get, and what

[B]

if

[C]

the state owns all the newspapers and all the radio stations and all the, you know, social media platforms now in our age, then they can set whatever policies they want because they're the owners.

[B]

Do you think criticism of the government is going to be welcomed and championed and cherished? Probably not. Okay, so that's a problem too, with outright central planning. That isn't a, quote, economic issue. But again, none of that stuff is in the critique of socialism that Mises originally put out. All right? So, and for me to distinguish among all these different flavors and, you know, even just qualitatively, what kinds of things could go wrong with outright central planning. It's. I'm not diminishing the importance of corruption or the incentive problem. I'm just clarifying and saying what Ludwig von Mises put his finger on, you know, first in an article and then later in a whole book, I think

[C]

it was, in 1920, devoted to socialism.

[B]

What he put his finger on was a completely independent problem. And it's Something that at first the socialists didn't even recognize. They denied that there was something there. And then only later, the more sophisticated socialist theorists, particularly those who had been trained in mathematical economics, came back and tried to answer him. And then in that, with that context, Hayek comes along and, you know, tags Mises and jumps into the ring, too. Okay, so that's kind of some of the background here. And I just want to explain. I know some people might hear this episode and think that this. That I'm somehow, like, attacking Hayek or

[C]

something, and I'm not. The stuff Hayek wrote on this is important. And to fully appreciate how impractical and

[B]

what a bad idea it would be

[C]

for your country to go down the road of outright central planning.

[B]

Reading the stuff Hayek wrote on this topic is very useful and gives you

[C]

more considerations and fills in some of the details. That's all true. However, I'll just say one more time and then, don't worry, folks, I'll end

[B]

this apology, as it were.

[C]

These large string of caveats, nonetheless, what people think of as the knowledge problem that Hayek identified, the problem of dispersed

[B]

knowledge, that is not what Mises was

[C]

talking about when he talked about the calculation problem.

[B]

Okay, so why am I talking about this now? The recent news hook was, for some reason, a 2023 string of tweets from Darren Asamuglu, who's a famous mainstream economist. He was at MIT back then. I think he still is. His book, he's got why Nations Fail is one of his big books. So anyway, he's very, well, public. There are jokes going around about how you could go your whole career as an economist and not even read all of Darren's stuff. All right, And I may, if I refer to him for the rest of the episode, I'm going to call him Darren, not because he and I are buddies, but because I don't want to mispronounce his last name in case I'm not saying it quite right. All right, so back in 2023, he had this long. Was it called a tweet thread on Hayek? All right, and he starts it out saying, this is a wonkish threat about information prices and decentralization. It touches on topics often discussed in economics, but with a revisionist take that might have relevance beyond economics, in particular about AI and control of information. All right, so then he talks about, oh, the most celebrated work from Friedrich Hayek is the Road to Serfdom, but then the next one, and this is the. The important tweet that made Me decide to devote this episode of the Human Action podcast to this topic is he says Hayek's argument offers an original and ingenious and then quotation marks computational critique of central planning. Sorry, I should have said in the prior tweet. He first holds up the road to

[C]

serfdom and says, this is Hayek's most celebrated work.

[B]

That's what most people know him for, but that's kind of polemical and for the man on the street, whereas his most important scholarly article in terms of among economic theorists was his journal article called the Use of Knowledge in Society.

[C]

Okay, so Darren had said that.

[B]

And then now the follow up, you know, the next in this thread, he said, Hayek's argument, meaning in that journal

[C]

article, offers an original and ingenious computational critique of central planning. His basic premise is that there is a huge amount of dispersed knowledge in society about a very large number of goods and services. And then in parentheses, example, people's preferences. Okay? And then know, Darren keep goes on and says, high guard, use the market system of prices, aggregates the information.

[B]

Da, da, da.

[C]

And then he says, Hayek posits that central planning wouldn't work because it would be impossible to collect and compute the right allocation of resources. Central planning is doomed because it fails to leverage this dispersed knowledge in society. All right, and then Darren goes on to say, this argument has always struck me as unsatisfactory.

[B]

And then finally, the thing that made this, you know, relevant again is he said a few. I'm, you know, jumping ahead here. He says, impossible to know the answer to this, but some believe that advances in AI are taking us toward this type of supercharged computational power. In my mind, this does not make central planning any more attractive. But lately I've also come to realize that there was another implicit step in Hayek's argument. Okay, and so he's going through that. And so his point was I, I can stop here because that's. We don't need to get into the weeds of what Darren's particular nuanced take was. I'm just explaining the context here. You know, he went to say, really, if you think about it, what Hayek ought to have been stressing is such and such, because if his argument really is what I just summarized, then it seems like modern AI solves or greatly mitigates that problem. Right? Okay. You know, we've got AI systems that help Amazon, you know, recommended for you, and they can look at your past purchases and figure out you might be interested in this. And then what if it gets to the point, some people are talking about this, what if it gets to the point where they, Amazon use it or whatever, you know, sister companies might be spawned in the future based on that model and you just, you know, open your door and drones have delivered to you some products and you open it up and you're like, oh yeah, yeah, I totally like this stuff, right? Like, if it gets good enough that it doesn't even have to show you first that it can just ship it to you. And so you don't have to wait. You know that. And if you're wondering what problem does that solve, you don't have to wait, right? In other words, if you're scrolling, it says recommended for you. And you, you agree with the algorithm that, oh yeah, you're right, I do want to read that book or I want to get that bath product or whatever, take that toy for my kid,

[C]

and yet click order. You still gotta wait a day depending on your, you know, if you got

[B]

Amazon prime or whatever to get the thing.

[C]

Okay, but if it got so good that it can anticipate what you want or what you will want, right, to

[B]

make it even more concrete that you don't even know you want it until you see it and then ship it

[C]

right to you, and that's when you first discover it. Okay, so Darren's point is, whether you think what I just said is going to come true or not, certainly we

[B]

can appreciate and acknowledge it's not impossible

[C]

that that would become true or that we could get close to that or closer to that in the coming years. And so then doesn't that sort of take the wind out of the sails of Hayek, who is saying, you know, leaning so heavily on the claim that, oh yeah, there's this dispersed knowledge and that's why central planning doesn't work. Because even if they have the best of intentions and want nothing but to help their comrades, the few experts, you know, running the government and coming up with five year plans to run the economy and so forth, how are you going to get all of that dispersed tacit knowledge into their minds when they're crafting the blueprints, right? Even if they go ask all the engineers and agronomists and things like this, tell us, you know, forward us all of your technical knowledge so that we have all the latest production possibility plans at our fingertips as we figure out how to deploy society's scarce resources. The point is you can't communicate all of that relevant knowledge into the minds of the planners, right? There's different Metaphors people have used in the Austrian tradition to kind of get this point across. Stuff like, you know how to ride a bike. Okay, what if your pen pals with somebody in another continent and there's like, ah, my parents just got me a bike. I don't know how to ride it. Can you explain it to me, you know, over text?

[B]

And it's like, no, you, you. That's not the kind of knowledge you can break down and communicate through that medium. I could, I could show you how to ride a bike and give you feedback when I watch you try to do it and whatever. But it's. It's not something I could just type up and tell you, this is how you ride a bike. Even though I know how to ride a bike. Right. So there's stuff like that. And then more relevant to economic issues, Things like if you go grab a very successful plant manager, you know, some guy who runs the plant that makes contact lenses or something for Georgia, and he's been doing that for 20 years and he's really good, and then you say, okay, write up everything you know and send it to the central planners. There's things that he knows and he. He knows, oh, yeah, in this situation, this is what you would do. And he doesn't even know that he knows it. Right. It wouldn't even occur to him to write that out. Even though in reality, if he were faced with that scenario, he would know what to do, but he wouldn't be able to tell someone else that ahead of time. Okay, so things like that. So, yes, when you're trying to explain why doesn't central planning work, that type of consideration is important. And then even of even more relevance to economics, what Hayek showed in the Use of Knowledge in Society and his other essays is how the price system allows everybody to act on the basis of the dispersed knowledge. Right. That he famously says, like, oh, if,

[C]

you know, some copper mine collapses somewhere,

[B]

and temporarily there's a relative scarcity of copper.

[C]

The price goes up, and then all

[B]

the people around the planet who use

[C]

copper cut back on their usage. And those who are able to cut back more because the thing they were using the copper for, they have closer substitutes and stuff like that. They will cut back more.

[B]

And the people is like, no, we

[C]

gotta have copper for what we're doing.

[B]

Well, they're gonna just keep doing it.

[C]

They're gonna pay the higher price and just suck it up.

[B]

Right?

[C]

And that's exactly what you want.

[B]

And the important thing he said is

[C]

the people don't need to Know, why

[B]

is copper more scarce right now? They just need to know that it is. And so Hayek argued that the price

[C]

system, in a very economical fashion, communicates the relevant signals to every. To all the people who have to make decisions.

[B]

Okay, so fair enough, but two points here. So one is, I want to say that is not the same thing as Mises calculation argument. All right? Mises calculation argument stipulated all of that. Mises said he concede. He concede a bunch of stuff. For the sake of argument, he said, number one, let's assume that the planners are noble, benevolent people, that they really are just trying to more efficiently organize their society's means of production in order to deploy them into the most socially beneficial channels, okay? In other words, they're not. They didn't seek that position because they love just holding arbitrary power over their fellow man, or they don't want to punish their enemies or, or, you know, they don't have territorial ambitions and want to seize control to, you know, beef up the military. No, imagine that. Let's just for the sake of argument, suppose there is an absolute angel who took over the central planning job. Okay? So Mises concedes their intentions. He also concedes the motivations of everybody else in society that, yeah, suppose that no matter what the, you know, planners suggest, there's no grumbling, there's no shirking. Everybody happily reports to what, you know, whatever job they relocate to, whatever city they live in, the most abhorrent conditions if the plan calls for it, perhaps temporarily, you know, yeah, we got. You got to move out of your nice house now and go live in this hut for a couple years because we need those resources to go do something else that's more important, you know, and they say, okay, yes, sir, yes, comrade, we happily do it for the good of the socialist commonwealth. Okay? So he assumes way all the incentive problems, and he also assumed away the concentration and aggregation of technical knowledge and consumer preferences. Mises even stipulates and says, suppose the central planners know perfectly well all the different technical recipes, and how do you take a given combination of inputs and

[C]

transform them into various possible combinations of outputs?

[B]

Suppose that there's no problem, you know,

[C]

there's no ignorance on the part of

[B]

the planners that they're drawing up the blueprints without access to that technical knowledge that they assume that they have at their disposal, you know, all the expertise that's embedded in all the people in their community. Okay, so he concedes that. And like I said, he also concedes

[C]

that they know how the public would rate various possible combinations of consumer goods, right? So to say, would it be better if we could make 16 more cartons of size 5 diapers for babies, or should we make 18 more television sets,

[B]

you know, of these dimensions and quality

[C]

and that suppose there's no doubt that the planners either can just use their own rankings on that, you know, being paternalist, or that somehow they have access and get a feel for what the public thinks in terms of ranking various

[B]

types of, you know, discrete bundles of

[C]

consumer goods and services. And Mises says, still, even if you stipulate all that, there is this fundamental

[B]

calculation problem that the planners, in order even ex post to say, did we come up with an efficient use of society's scarce resources? That we told everybody, okay, here's what you do. Use the farmland and you're going to do this, and here's the buildings and this many workers are going to report to each building and work this many shifts and blah, blah, blah, and you're going to get this much steel and you're going to produce this many cars and you're going to produce this many decks for homes and these locations and grocery stores are going to be located in these coordinates and blah, blah, blah, blah, right? You're going to do all that stuff and everyone does it. You know, nobody shirks, nobody commits private crime on the side, right? All this stuff. Still, the problem is nobody. Would know, ex post, whether that constituted inefficient use of society's scarce resources because the people, they don't have the costs, right? They don't know the value of the inputs that were used up compared to the value of the outputs that were produced. That's the fundamental problem. So let me just mention at this point that I will put in the Show Notes page a post that I wrote for Mises.org a few years ago at this point where I really carefully go through and spell out how there's a distinction between the massesi and calculation argument in the Hayekian knowledge problem or assessing calculation problem. Let me also, just as a, an offshoot of that, explain that. You know, I don't think that I'm quibbling here that the market socialists themselves, like the economists who back in the day, during what we call the socialist calculation debate, argued with, you know, against Mises and then Hayek when he joined the fray. They also perceived a difference rhetorically. All right, so let me just. So here I'm cribbing from my mises.org piece, which again, we'll, we'll link to in the show notes page. But here. So H.D. dickinson in 1933 is responding to Mises, you know, original salvo. Right. So it. And by the way, Mises, it was in 1920 that he had an article that came out, it was in German originally, where he lays out what we now call, you know, the calculation problem or the calculation argument against socialism, where Mises says what? I just walked through that. Without having a common denominator, you can't. Without being able to put all the various types of inputs which originally are just, you know, different types of goods and services and things. You know, how do you amalgamate 16 hours of a brain surgeon's time and 16 hours of a street sweeper's time and six barrels of crude oil? Like, how do you add those all together and say, what's the value of those resources? And then compare it to the output that you might make and say, oh, well, you know, we could use those to. With some other things too. Obviously we could contribute and make whatever this many diapers and this many jet aircraft engine and this many did it. And how do you, how do you say, was it worth it? Did we produce more valuable outputs than the inputs that were used up in producing them? And again, if originally in the raw state or their original state, these things are just different units of various goods and services and, you know, resources and things like that. How can you do it? It's not. The point is, it's not an engineering problem. It's not an issue of getting the right expert in the room. All right, now where it is solved in practice is in the market economy, where those resources are privately owned and there's genuine market prices that are formed. And so then, you know, ex post, an accountant can come in and show the owner, Yep, here's the profitability of the various lines in your enterprise. Yep. And these, these lines were profitable. These ones weren't. And what do we mean by that to say, oh, the revenue you got from your customers was insufficient to cover the expenditures for those lines? That's what it means to be unprofitable. You suffered a loss. Okay, and I'm, I'm dumbing it down a little bit, but that's, that's the gist of the argument. So that's the problem with outright central planning, because there isn't genuine private property in markets and genuine market prices for the means of production. And then Mises pivots to show, and this is how capitalism, you know, the market economy with private property and the use of money, that's how it solves this problem that plagues socialism. Right. Okay, so Mises laid all that out originally in 1920 article, and then he later had a whole book on socialism that came out shortly thereafter. So Dickinson in 1933. Used an argument based on Walrasian general equilibrium. Right. So. So Dickinson was a trained mathematical economist, and he was saying, look at the socialist. Central planners can solve things in the same way that the market economy does. Right. And Dickinson's understanding of how the market economy solves these problems is, oh, here's a system of simultaneous equations, and you've got all the consumer preferences, and you've got all the production functions. And so it's a matter of just, you know, finding a price vector that satisfies all these equations where the firms maximize profits and the households maximize utility. Blah, blah, blah, blah. There you go. And this is how we do it when we're modeling a market economy. You know, if you went and studied economics at the top schools, that's what they would teach. And, you know, how does the market work? That's what you would do. How do you find the competitive equilibrium? There you go. And why couldn't central planners at least approximate that? Yeah, maybe they'd get something wrong here or there. But this Mises guy is making it sound like qualitatively there's some, you know, fatal flaw with socialism when. No, what are you talking about? This is what they would do. Right, that. So that's what his argument was. And so Hayek answered that in 1935, and here's what. What Hayek said. So he referred to these mathematical solutions to the Mossesian challenge, and Hayek said, yes, what Dickinson wrote and then quote. This is the exact quote from Hayek is not an impossibility in the sense that it is logically contradictory. But then he went on to say, the reason, you know, Dickinson's attempted response to Mises doesn't work, and this is hock. Now, what is practically relevant here is not the formal structure of the system, but the nature and amount of concrete information required if a numerical solution is to be attempted and the magnitude of the task which this numerical solution must involve in any modern community. Okay, and so then why. I think it's, you know, so I'm saying, just looking at that for number one, you can see where Darren's coming from. And he's saying, you know, I think Darren did a good job of summarizing what Hayek's point was, by the way. The. The use of knowledge in society came out, I think, in 45. Right. So Hayek has these famous papers called the Knowledge Papers, and they came out after Hayek, you know, crossed swords with the socialists. So in the standard telling of this stuff, it's like, oh, Mises and Hayek didn't even realize how much their fellow economists had diverged from what they thought were sound economics. In other words, if you read the early Mises, you can see him saying, I don't remember the exact phrasing, but kind of like, oh, yeah, all the, all the major economists now in the wake of the subjective marginal revolution, we all agree on these important matters. And that early in his career, Mises has some passages where it looks like he thinks all the actual, you know, professional economists at this point are all basically on the same page, which might be surprising. And I, and so I've heard some argue that it was the socialist calculation debate that made Mises and Hayek realize, oh, wow, these guys don't think like us at all. Like, they really believe in these mathematical models of general equilibrium theory and da, da, da, da. And so after having gone through that gauntlet, then Hayek later writes what are called the Knowledge Papers, including the use of knowledge in society. Right. So that's, you know, I've, I've heard people teach, like, the history of economic thought from an Austrian perspective with that narrative. And that rings true to me. I think that's, that's probably right. Okay. So anyway, in 1936, Oscar Lange joins the fray. Right? So 33 is Dickinson. 35. Hayek, you know, wrote his response that I just went through. And then a year later, Oscar Lange comes out. So Lang is another socialist. And he opens his article by paying mock homage to Mises, you know, saying, yes, previous socialist theorists thought we could just get rid of money. That that was just a barbarous relic from the, you know, corrupt capitalist society. But thanks in part to, you know, the, the objections leveled by this Ludwig von Mises character, you know, we cutting edge socialist theorists now acknowledge and recognize, oh, you do need a price system. We, we can't just get rid of that. That, that is important, but we don't need private property. Right. In other words, Mises, Lange of thought

[C]

was

[B]

putting too much weight on the private property element. And Lange just said, no, no, you definitely need prices, but they're really just official, you know, exchange ratios that they can just be announced by the central planners. You don't need a decentralized private property framework with various people bidding on iron ore and farmland and whatnot, that you can have the central Planners just announce vectors of official exchange ratios, and we'll call those prices. And then we just give orders to the factory managers, go pretend that you're in a capitalist system and try to maximize profit on the basis of these exchange ratios that we're going to announce. Okay, so that's where Lang is coming from. But he said, thank you, Mises, you know, and he even says, I think that the socialist commonwealth of the future should have a statue of Mises to commemorate, you know, this contribution he unwittingly made to our noble, you know, conquest of Earth through our superior wisdom and planning and such. Okay, so after he pays this mock homage to Mises, Lange then says, hey, here's what Mises originally said. Then guys like Dickinson gave the response and said, what are you talking about? We use mainstream mathematical economics to solve these problems that Mises raised. And then look what Hayek just said. And he, you know, quoted what. What I just read to you folks. And so then, now here's what Lange has to say about that. Thus, Professor Hayek and Professor Robbins there. He's talking about Lionel Robbins, have given up the essential point of Professor Mises position and retreated to a second line of defense. On principle, they admit the problem is soluble, but it is to be doubted whether in a socialist community, it can be solved by a simple method of trial and error, as it is solved in the capitalist economy. Okay, so again, the important point here is rhetorically, Lange was saying, Whoa, whoa, whoa. Originally, Mises was throwing down the gauntlet and saying, economic calculation in a socialist framework where there's no private property in the means of production is truly impossible. Like. Like, logically speaking. But now, you know, the mathematical economists who happen to be socialists, we're showing. What are you talking about? This is how you could do it, at least in principle. And then Hayek says, well, no, no, no. Yeah, sure, it's not logically contradictory, but I mean, come on, in reality, look at how complicated a modern society is. Do you understand what would be involved, you know, in contrast to your simple, you know, models that while Ross would give you with, you know, an apple, apples and oranges, and two people trading, and you draw your indifference curves and blah, blah, blah, Right? That. No, a modern economy with all the goods and services they have and all the dispersed knowledge. Come on, you couldn't do that for real, not in practice. And so Lange is going, aha. See, that's a retreat. Mises didn't just say it would be unwieldy or that the market economy could solve these problems more rapidly or approximate the answers more closely. No, Mises was saying in principle, socialism can't do it, period. Where's the market economy, Ken? And now you, Hayek, haha, have weakened the claim to just saying in practice this couldn't be done, but in principle it could be right. So I don't think Lange is being slippery there. I think what he said is true. Okay. And you know, now maybe someone who's a fan of Hayek might say, well, right, because if you cast Mises original claim in that really dogmatic sense that he is wrong. All right. And I, I, I have seen people summarize Mises argument in a way that would render it to be invalid, put it that way. Right. Like I've seen some people say, yeah, Mises was conceding that the planners could, they knew how to all the various combinations of, of consumer goods and how to rank them. And they, the planners also know every possible combination of outputs that could be produced given the original resources. And so, but they wouldn't know how to do it. And if you're familiar with mainstream economics, you can say, okay, but then that means you kind of just conceded the indifference curves ranking all the possible output bundles and the production possibility frontier showing all the combinations of output that we could produce. And so don't you just draw those two things on a graph and then find the point of tangency and you know, so you can do that, at least in principle. What are we talking about? Right. So anyway, but my point is whether you think Hayek wisely retreated strategically to a more defensible fortress or whether you think he foolishly, you know, agreed to too much and threw out the bulk of Mises argument and you know, try to define, defend a weaker claim. Either way, I think it's crystal clear that Hayek did retreat from what Mies original claim was. And so it's again, I don't think it's just unnecessarily quibbling or something or just trying to cause controversy for some of us, particularly in the Rothbardian tradition to say the Messesian calculation problem is distinct from the Hayekian knowledge problem. Okay? So the way I like the way Joe Salerno handles this, it I had all these things going around in my mind and then when I read this article by Joe, it helped crystallize for me and say yes, that that's what it is and it helped me move on with my life. I, I, I resolved the, the tension that had been building. Okay, so in a 94 article Joe says I conceive appraisement. So what appraisement means is that an entrepreneur is looking at various inputs like, you know, capital goods and, you know, raw natural resources, perhaps some workers with their various capabilities in terms of labor output. And the entrepreneur appraises what is the value of this bundle of inputs? Okay, Meaning ultimately like, what could I sell this stuff for? Or if I what could I produce with this bundle of inputs and then sell that for? And so you're going through that mental process to appraise, to assign values, you know, monetary values to this collection of items. And so here's what Salerno says about that. I conceive appraisement is neither knowledge nor arithmetic, but is something new under the sun, introduced into the world only when the institutional prerequisites of a market economy are fulfilled. The social process of appraising thus transcends the purely individual operations of knowing and computing at the same time that it complements them in creating the indispensable conditions for rational choosing by entrepreneurs and resource owners cooperating in the division of labor. Okay, so here, let me just because I thought this was a good analogy I had in case you're still struggling. You're like, well, yeah, but I still don't get it, Bob. So I'm still not quoting from my mises.org article. Let me try to motivate the point this way. When we use a thermometer to measure the temperature inside an oven in a commercial bakery, the device transmits information to us. There really is an objective fact of the matter of the kinetic energy of the air molecules bouncing around inside the oven. And the thermometer is an imperfect way of translating that data to us in a form our minds can comprehend and incorporate into our decisions. But there's no doubt that the oven really has a temperature regardless of our measuring it with a thermometer. In contrast, if we ask, how much economic value does this oven possess, then that's a fundamentally different question. This isn't an objective fact that is embedded in the arrangement of matter. The question takes into account all of the subjective preferences of everyone on the planet Earth, as well as their expectations about the possibility of transforming matter into different forms. It is a mind boggling question, in fact, that can be answered only in a by setting up a market economy and then making informed guesses as to what people would be willing to pay for that oven. Okay, so I hope that helps. Again, whether you think this is worth me spending an episode on or not, I hope you I've driven home to you and shown the distinction. Right. That, in other words, it's the way the market socialists were thinking about it. And I think this was endorsed by Hayek, and the way he was arguing with them is that, yeah, there's a bunch of dispersed knowledge in different minds around the economy. And the problem, the central planners is just if they could somehow mobilize all that and get access to all that information, that raw data into a centralized location where a decision maker could have access to it in real time when making decisions, but that the form in which this. This raw information exists is fine, and it's just a matter of getting all that stuff into one spot. And what, you know, Joe is arguing is that, no, that that's missing something that when we talk about economic calculation. And the thing that Mises was putting his finger on when he said the central planners would lack this, it wasn't just that. Oh, yeah, because the people who build bridges have a bunch of localized, specialized knowledge that even in principle it would be, they couldn't transmit to the planners. And then these people over here understand that, you know, oh, yeah, sometimes when I want to go out to eat, I'm in the mood for something sweet and spicy, and other times I want something savory, and other times I want something, you know, blah, blah, blah. And it'd be hard for me to write that all up and get it across to somebody who's making decisions on my behalf. I just kind of need to be in the moment. And, yeah, these are all true statements, and they're important to know and to understand when you're grappling with, you know, different arrangements for economic organization, if you want to use phrases like that. But the point is, that's not what Mises was talking about. He was talking about something else, something qualitatively more fundamental. Okay, all right, so namely that you need this institutional setup of decentralized private property ownership and the existence of a commonly accepted medium of exchange, such that in the haggling in the markets for the various resources, genuine market prices emerge, and then you finally have a common denominator, and you can say, you know, using expectations. And maybe you're wrong, but the point is, the thing that you're trying to forecast is the right type of thing. Whereas if you're just a central planner, it's not enough just for you to kind of look ahead and say, well, gee, would people want more, you know, SUVs in this scenario? Or maybe under what circumstances would they prefer to have the bridge? And what, you know, maybe we make More diapers, maybe we make more turtlenecks, maybe we have more sporting events. And that pulls, you know, labor out of certain other potential occupations. But the people like watching sporting events, but some of them don't. Geez, what do we do? Right? That it's not merely a technological engineering, computational problem, that it's beyond that there's something more fundamental at work, that the type of information that market prices convey is different. Maybe even what I just said, there's not quite. Right. Like the market price is the information. It's not that. It's just, you know, a surrogate or a proxy for something else. I think that's part of the issue that people are trying to get across here. The ones that are really stressing and saying, whoa, don't just read some of Hayek stuff and think that he accurately summarized Mises point. Okay, okay, so now I want to throw all that and let me now circle back to the original thing. The reason I don't think I said this, the reason I'm talking about this, he said, you went back three years, Bob. And this thing that this Darren guy talked about, three years, if for some reason it was making the rounds again, I don't think I said that at the outset. So that's why it's topical again. And people just recently were arguing about this on social media. And so here, let me give what my response is that, Right, if you, if you conceived of the problem with socialism to merely be one of dispersed knowledge, and particularly if one of the big components is just, oh, you don't know what people are going to want, then AI could help you deal with that. That's true. That would help mitigate the problem and make your forecasts more accurate. Okay, that, that's, I think that's an unobjectionable claim, but there's two problems with that. So number one, again, I think what that mischaracterizes what Mises objection to socialism was, maybe it's a fair characterization of Hayek. And by the way, I saw some people arguing and saying, well, no, it's, you know, this doesn't solve the problem because there's tacit knowledge and stuff that you can't communicate. I mean, yes, but AI can help solve that. Right? Like, I sometimes the AI algorithms, you know, that are employed on Amazon or whatever, I think they do a pretty good job. And I think they're clearly going to get better over time. And so I don't know that there's this qualitative gap between, you know, the types of things that Hayek was talking about and what AI could eventually do. Right. But there does seem to be, I think, a qualitative gap between that and what Mises was. Was talking about. It put it to you this way. The only way that the AI systems could approximate the problem is they would have to simulate. Like if. If all planet Earth were socialist. That's what we're talking about. Could a bunch of AI systems being used by the central planners, could that help them overcome the necessity and calculation problem? I say the. The only thing they could do is try to model and say, what if right now we had planet Earth with all of us starting, you know, natural resources, all the different labor skills that people have, the location of everything, because that's important too. Where a factory is located is a big part of how do we integrate it into, you know, the blueprint for what's the economic plan for the next five years? It matters where all the factories are located. If they're all located in Idaho, that means one thing. If they're spread out uniformly around the globe, that means something else, right? So that's, you know, plug all this raw data in as inputs, and then if you had the AIs did a simulation and said, now imagine there's private property in the means of production, and then we generate market prices and, you know, come up with, like, a simulation of a market economy on this, you know, alt earth, this alt capitalist earth. And then the central planners looked at the outcomes in that simulation and said, oh, so a ton of steel trades for this much against those, those, you know, farmland of this type of, you know, air ability and whatnot. Oh, okay, fair enough. Then that might help them some. But again, it wouldn't be simply aggregating, you know, consumer preferences and things like that. Like, you would have to be simulating to try to say, if we had a market, what would it do? And then you could argue, like, the better that simulation is, you know, the word, the. The more you'd be containing the errors of the socialist decisions. Okay, but it would have to be something like that, which I don't think is what Darren was getting into. Okay. However, even having said that, my next and final point here is to say, even thus far, still, at best, all that would show is central planners in the year 2026, using modern AI, could try to approximate what the market economy could have produced spontaneously in a decentralized fashion based on, you know, the year 2000 technology with fewer computers. Okay? Because all the computers needed and the AI systems that were dedicated to simulating the market economy to give the central planner some guidance as to the, you know, the relative trade off, the relative valuation of various types of inputs, those, you know, supercomputers and AI systems, whatnot. If we switch to a genuine market economy instead of having a global socialist commonwealth, they could be used by industry in order to enhance what it is that the private businesses did. Okay, so again, given that the decentralized price system and private property and so forth in the existence of genuine market prices and the aid they provide to calculation in the form of just accounting operations, given that that contributes something to humanity and that you're trying to approximate that with some make arounds using supercomputers and AI and so forth under socialist framework, I'm saying you, you have to take into account the opportunity cost. You can't ignore that. And so even if you thought the AI could give a decent approximation such that the calculation problem wouldn't be fatal, again, I'm not conceding that, but I'm just saying suppose that were the case still you're using, you know, a large amount of society's potential resources for this task that the market economy would have solved. Maybe effortlessly is not the right word. I mean he takes accountants and things and you could argue that, you know, under socialism the accountants, the human accountants

[C]

get freed up to do something else.

[B]

Okay, but that's, I would just point that element out and I, a lot of times don't see people explicitly accounting for that, no pun intended, when they talk about doesn't AI solve these problems? That it's not put you this way. I think some people believe that. Oh right now if Amazon's using AI to help determine, you know, oh, Jim Smith here in, in Tallahassee, he's ordered such and such in the past and so why don't we send him these books and why don't we send him these sneakers and this stuff or whatever up and it's. We know it's his daughter's birthday coming up because it is every year and he, he typically buys this kind of stuff but you know, sometimes he's late and he has to do a rush order. Why don't we send it a week ahead of time? And I bet you his daughter will like this stuff and probably will even he'll like it even more to give to her than if he went browsing on his own last minute. Suppose that's all true. That doesn't prove. Oh, so actually socialism could get by and that, you know, the Austrians, maybe they had a point back in the 30s. But now modern technology has rendered their objections obsolete. No, because still the AI helps Amazon be more valuable and, you know, more profitable and provide happiness to its customers embedded in a capitalist framework. Right. You, it's still the case you can't send the guy in Tallahassee more stuff than he's able to afford. That wouldn't be efficient. That would be a waste of resources. Okay, so I'm saying all these advances that companies are making right now in the real world with AI, those improvements are only able to be steered properly and to actually be focused on where they're going to provide genuine value if their owners are embedded in a market economy and are responding to profit and loss signals. Okay, I will stop there. One last bit I also have, for those of you who are really wonkish, I have an article in the QJ on what's called Cantor's Diagonal argument. And that was something I wrote years ago. Weighing in on this issue of oh, is so is socialist calculation merely impractical or is it genuinely impossible, even in principle? And I waited on that. And so if you're familiar from mathematics, loosely speaking, there's a sense in which the real numbers between 0 and 1, that that's like a bigger set, it's got a higher cardinality than the integers, you know, 1, 2, 3, 4, the positive integers 1, 2, 3, 4,. All the way up quote to infinity. That there's a sense in which there's like the density of numbers in between 0 and 1 when you can include decimal places is bigger than the set of integers. And I use that fact from math to then circle back and make the argument that no, even in principle, you could never mimic the market economy with socialist planners who are using the mathematical approach from certain branches of theoretical economics. Alright, so if that sounds intriguing to you, I'll link to that too in the Show Notes page. Okay folks, thanks for your attention. See you next time.

[A]

Check back next week for a new episode of the Human Action Podcast. In the meantime, you can find more content like this on misa's dot org.