Key Takeaways

  • Bhaskar Sunkara, founding editor of Jacobin and president of The Nation, explicitly sides with Austrian economists Ludwig von Mises and Friedrich Hayek over central planners in the socialist calculation debate.
  • Odd Lots co-host Joe Weisenthal argues that sufficiently advanced artificial intelligence could turn entrepreneurial hypothesis testing into an algorithm, removing the need for private firms.
  • Sunkara counters that software can calculate static technical allocations, such as laying railroad track, but cannot capture shifting human preferences or resolve subjective trade-offs.
  • In his book The Blueprint, Sunkara outlines market socialism: keeping market-driven prices and consumer choice while replacing wage labor with worker-owned cooperatives, public banking, and socialized land.

The Disagreement

Joe Weisenthal suggests that artificial intelligence could finally resolve the core hurdle of central planning. In his framing, the modern capitalist firm exists only to test guesses about what goods people want. As Weisenthal puts it: "And if you perceive capitalism as a series of entrepreneurs making hypotheses and then testing them with the firm, I actually, not only do I think it's possible, I think there's a very good chance that AI leads to essentially a form of communism by which we no longer need the firm as a hypothesis of what will work." If an algorithm can predict demand and simulate experiments, the private firm loses its purpose.

Bhaskar Sunkara rejects this techno-utopian conclusion. Even as an avowed socialist, Sunkara refuses to treat price signals as an artifact waiting for algorithmic replacement. He points out that “market socialism is now the norm on the socialist left,” but warns against comrades who assume that “in the future, AI will solve calculation problems.” Instead, Sunkara takes a stance that shocks orthodox Marxists: “I'm on the von Mises/Hayek side of the calculation debate, and and I I also believe in the inherent good of the market as a coordinating mechanism.”

To Sunkara, technical optimization is entirely separate from economic value. A central computer can calculate the materials needed for a project, but it cannot determine whether society wants that project compared to thousands of alternatives. As Sunkara explains, “You could answer the calculation problem of like the most efficient way for the state to build like a railroad, and like that the the correct allocation. But the reason why we have like market-driven prices is obviously for the trade-off end of it.”

Who's Right (and When They're Wrong)

Weisenthal is right only when a problem has fixed boundaries and static variables. If a business runs a closed logistics network with known inputs, predictive models can allocate freight better than human dispatchers.

Where Weisenthal is wrong, and where Sunkara's defense of Austrian economics wins, is assuming human desires are fixed data points. People do not carry around a tidy list of preferences waiting for a neural net to scrape them. Preferences reveal themselves only when buyers face real scarcity and part with hard-earned cash. Price signals transmit that dispersed knowledge faster than any model can simulate it.

Execution breaks central computing even faster. A plan is obsolete the minute physical reality intervenes. Sunkara highlights this practical flaw: “Like I think I think AI can help form a better plan, but it wouldn't actually help you execute that plan. And then if something changes, you would then have to reform the plan. It's kind of would be the worst of both worlds in many ways.” When reality conflicts with the forecast, a central algorithm must rewrite the entire system. Independent firms adapt immediately because local survival depends on responding to price signals.

What to Do With This

Stop using synthetic personas or predictive software to decide what features your users want. Change your pricing page tomorrow morning, put an actual dollar figure on your new tier, and pitch it live to your next five customer calls. Real willingness to spend money reveals market truth that no algorithm can generate.