It has been almost a year since OpenAI unleashed its ChatGPT program. Every day since the business press has been awash with stories of its effect, and the effects of AI in general. For instance, investors poured $40 billion worth of venture capital into AL start-ups in the first half of 2023, five times the amount of last year. This is despite a 48 percent plunge in investing overall. The number includes the $10 billion Microsoft pumped into OpenAI with the goal of Microsoft’s Bing Chat disrupting the search engine and targeted advertising world Google has dominated for decades and incorporating AI features into its Microsoft Office software. Google allegedly declared a ‘code-red’ to catch up while being criticized for getting caught off-guard and being too cautious in its AI approach- though at this point Google still handles about 90 percent of search queries (generating over $160 billion in annual advertising sales). Google added AI text generation to its search engine and now offers its own chatbot, Bard. Meanwhile, Nvidia became the fourth company to reach a trillion dollars in valuation. The company makes up about 60 to 70 percent of the global supply of AI server chips. Amazon recently announced a $4 billion investment in OpenAI rival Anthropic which has created a rival to ChatGPT called Claude. Amazon is also seeking to make its Inferentia and Trainium chips rivals to Nvidia’s.

Beyond the purely economic, ChatGPT has triggered a society-wide debate and panic about what the development of general AI means for human endeavors from education to art to warfare. The writers’ and screen actors’ unions strike that has ground much of the TV and film industries to a halt has the use of AI as one of its core issues. This past March hundreds of scientists and leaders in the field of AI called for a six-month pause in development so that society could gather itself since large language models were developing faster than expected. Will we inevitably submit to our computer overlords? Or ultimately be destroyed by them?

However, in the midst of all this, there is an interesting topic that has seemingly been overlooked by the mainstream. Capitalism has always had a chorus of sirens that have striven to allure Homo sapiens from thinking they had any chance to escape its clutches. Margaret Thatcher proclaimed ‘there is no alternative’ in 1980. In the wake of the Soviet Union’s collapse, Francis Fukuyama declared the ‘End of History’ in 1992. Given the fall of the Soviet Union, the decline and isolation of the Cuban economy, and the dominance of the ‘Washington Consensus’, the decade after the Cold War was one of capitalist triumphalism. It was during that time the Marxist literary critic Frederic Jameson famously wrote that ‘it is easier to imagine the end of the world than the end of capitalism.’

The late Mark Fisher described it as ‘capitalist realism’ (the title of his 2008 book on the subject), ‘the widespread sense that not only is capitalism the only viable political and economic system, but also that it is now impossible to even imagine a coherent alternative to it.’ Central to capitalist realism is the idea that an economy based on planning and democracy is not viable, inevitably leading to endemic shortages, bureaucracy, and stagnant growth.

Such were the arguments put forward by Austrian School economist Ludwig von Mises in his seminal 1920 essay titled ‘Economic Calculation in the Socialist Commonwealth.’ Mises asked how could planning boards know which products to produce, how much should be produced at a given time, which raw materials had to be used, and how much of them? Where should production be located and which production process is the most efficient? And how would all this information be gathered and calculated and then retransmitted back to all the relevant actors throughout the economy? Mises’ answer was that no human process could accomplish it. He argued it is simply beyond the capacity of any planning agency to accurately describe supply and demand across all economic sectors therefore planners working with flawed data would regularly produce vast mismatches between what is demanded and what is supplied, resulting in inevitable shortages and the requisite barbarism.

Instead, Mises argued the simple mechanism of prices floating in a market contains all the needed information. This argument was later taken up by Friedrich von Hayek. Hayek also viewed prices as information-gathering machines reflecting the discrete bits of knowledge scattered among executives, workers, and consumers. Prices, derived from the collective wisdom of the crowd, could coordinate information through a decentralized network Hayek called a ‘spontaneous order’, making planning unnecessary.

 While this argument was at least largely valid in the past, the rise of the bureaucracy and stagnation in the Soviet Union demonstrates this, does it still have the same power now? For instance, when the planning boards Mises dreaded can be augmented, even eventually replaced, by algorithms and AI, can we shift toward an economy that begins to move beyond capitalism?

As brilliantly described by Leigh Phillips and Michal Rozworski in their book The People’s Republic of Wal-Mart, large, successful enterprises, even while operating within a general market economy, do a great deal of large-scale planning internally. And these enterprises have been using forms of AI for quite a while. Some of these companies have larger market caps than most countries’ GDP. Apple and Amazon are worth more than 90 percent of the world’s countries. In 1970 the GDP of the Soviet Union, the second-largest economy in the world at the time, came in at around $433.4 billion. In 2021 Wal-Mart’s revenue was $572.8 billion. These organizations eschew internal markets. The different departments, stores, and suppliers don’t compete with each other. Everything is coordinated. To that extent one can say much of the global economy is already planned.

In fact, there is a recent example of a corporation that actually took markets seriously enough to attempt to incorporate them internally. In February 2013, Edward Lampert, founder of the hedge fund ESL Investments, took over as the CEO of Sears Holdings (the parent company of Kmart and Sears formed after the former brought the latter), one of Wal-Mart’s main competitors. Sears goes back to 1892. Its catalog once revolutionized shopping for Americans, particularly the many back then who lived in rural areas. Lampert announced his intention to create markets within the company, breaking it up into 30, then later 40, autonomous units that would compete with each other. Each unit had its own president, board of directors, and chief operating officer, and separately measured their own profit and losses. The idea is that this would efficiently produce better data.

Instead, it devolved into absurdity. Creating internal divisions blocked internal synergy. If a division needed help from the HR or IT departments, it had to write a formal request or use a contractor. In order to optimize profits at one division at the expense of others, infighting erupted over everything from floor shelving to advertising space on circulars. The results quickly spoke for themselves. A Bloomberg expose from 2013 described the gross spectacle of screwdrivers being advertised next to lingerie. Little funding went to needed upgrades at stores, many of which became dilapidated. Sales dropped by $10 billion. By October 2018 Sears Holdings filed for bankruptcy and Lampert stepped down as CEO (though he remained chairman). While Sears was facing stagnation since the 1990s as online retail took off, it was an epic Randian failure that truly crashed it.

Compare all that to the fluidity of Amazon. Amazon is certainly a soul-sucking corporation that grinds workers to dust. Yet it has achieved logistical and operational genius. Consider that at any given moment Amazon has 600 million items up for sale, basically all available to be home delivered within two days from strategically placed distribution centers that more and more run on algorithms and robotics. Amazon uses search and point-of-sales data and search history to stock the centers. The result: Amazon receives about 115 orders, basically a full delivery truck worth, every second. That’s 10 million fulfilled orders in a day. An estimated 60 percent of U.S. Adults are Amazon Prime members.

In a November 2019 profile for The Atlantic of Amazon founder and then CEO Jeff Bezos, Franklin Foer had this astute observation:

Amazon, however, has acquired the God’s-eye view of the economy that Hayek never imagined any single entity could hope to achieve. At any moment, its website has more than 600 million items for sale and more than 3 million vendors selling them. With its history of past purchases,  it has collected the world’s most comprehensive catalog of consumer desire, which allows it to anticipate both individual and collective needs. 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.

This last point is nonsense. Simply nationalizing Amazon wouldn’t achieve too much and in fact risks replacing the dictatorship of capital with another dictatorship. But the greater point holds. Such efficiency, flexible planning, and logistical power could be captured and used to create a just, egalitarian society. In a world full of crisscrossing cables, instant global communication, along with ever-expanding AI, the arguments of Mises and Hayek truly lose their power. There are now many trillions of pieces worth of data that could be used to make nonmarket decisions about how to allocate the use of resources. ‘Big Data’ understandably has a bad name among many leftists, however, data is the lifeblood of any planned economy. Rather than being used for surveillance and targeted advertising, it can be used to determine and fulfill peoples’ needs.

We have a rudimentary example of how this could work from Chile’s socialist experiment in the early 1970s. By the end of 1971, the Allende government had nationalized more than 150 enterprises, including twelve of the twenty largest companies in the country. Recognizing the difficulty of reordering the economy in the face of fierce opposition and American sanctions, the government instituted Project Cybersyn. The aim, using the limited computing power that was available to Chile at the time (there was only one mainframe IBM 360/50 available for the project, it relied instead on a network of telex machines), to connect data from the factory floor and the State Development Corporation in order to enable quick decision making in response to changing conditions. The system would provide daily access to production data and modeling tools the state could use to predict future economic behavior. A futuristic control room would facilitate communication and data analysis.

As described by Eden Medina in her book Cybernetic Revolutionaries, though primitive and ultimately not completed, the system did enable the government to overcome a general strike called by the opposition (and funded in part by the CIA) in October 1972. A command center was established in the presidential palace, connected by telex machines to operating units focusing on different sectors such as energy, transport, and banking. Shortages were quickly reported by minute-by-minute reports from the ground through the network allowing different enterprises to shift resources. Government data showed raw materials continued to flow to 95 percent of economically crucial enterprises and food supplies were maintained at 50 to 70 percent.

Project Cybersyn didn’t survive the Pinochet coup in 1973 so its full potential wasn’t tapped, yet the promise remains. It is easy to imagine what can be planned with today’s computing power and mountains of data. In a recent interview with Wired, Uber CEO Dara Khoscrowshahi explained ‘AI is part of the Uber DNA. We use large models to predict your ETA, to process documents that drivers upload, to predict your next order on UberEats, or to predict whether someone wants an UberX or Comfort, Balack or Electric.’ Having a much greater idea of what to produce and how much, it does not appear to be a huge leap to begin working toward decommodifying the needs of society in general.

Of course, it takes more than advanced computer modeling and AI to build socialism. It first takes the working class to democratically organize the means of production. This can ultimately only be won at the barricades. However, as examples from the technologies of the iPhone to Amazon to the COVID-19 vaccines show, planning works.

While there is a wide range of opinions as to when its beginning should be marked, there is now an emerging consensus that the planet is in a new period of geological history, the Anthropocene, one in which human civilization essentially creates its own environment. This concept no doubt causes many to tremble in fear but denial of our collective responsibility will not change it. The specter of global warming, the rise of AI, possible future pandemics, and other environmental challenges are awesome, but so are the possibilities of maximizing human freedom, ending war and poverty, and probing deep space. We cannot trust the irrational, unplanned market system with its destructive incentives to fulfill our potential. As the world witnessed with the COVID pandemic, far from the picturesque visions of Mises and Hayek, a reliance on markets leads to inefficiency, hoarding, and reactionary nationalism. The only good Anthropocene is socialist, its vehicle is an empowered global working class. It still has a world to win.

---