Earlier this month, Business Insider reported that Amazon's Just Walk Out technology was not run by robots, but by the eyes of "1,000 workers in India who review what you pick up, set down, and walk out of its stores with." And so what looked like a new trick was in fact old hat. The work of American cashiers and attendants (a high-income society) had simply been, like service-related jobs, offshored to India (a low-income society). And this transference of services from one economic zone to another is only about one thing: wage arbitrage.
The Seattle-based heterodox economist Alan Harvey put it this way in his excellent little book Demand Side Economics: Demand Side Minds.
Explicit in the new globalization is the free flow of capital and the opening and integration of markets. And while "trade" denotes an exchange, the current phenomenon is one of arbitrage of labor, regulation, currencies and financial instruments. Arbitrage is taking advantage of the price differences between two or more markets.
Amazon recently admitted that it does use humans in this technology, but only to train AI. In the near future, the tech corporation promises, humans will be completely replaced by robots. Nevertheless, Amazon is removing the Just Walk Out technology from its Amazon Fresh stores and replacing it with the Dash Cart, which is, in essence, a modification of the self-checkout kiosks found in most grocery stores. (Self-checkout, like the Dash Cart, transfers the paid labor of a cashier to the unpaid labor of a shopper—an extreme form of wage arbitrage.)
Nevertheless, the human-to-AI story does not add up. The problem, as explained to me by an Amazon Web Services (AWS) worker, who, of course, cannot be named, is Just Walk Out demands too many high-end cameras, relentless processing, and the constant movement of a vast amount of information. Therefore the system only works in places that have access to fast internet, which is not cheap.
Think of it this way: A change of mind has a price in an Amazon Fresh or Amazon Go store. This price does not exist in a conventional grocery. In the former, when a shopper decides not to buy something and returns it to the shelf, a massive load of information must be transmitted to a human or a robot about this simple and common shopping behavior. With the latter, no such expenditure (either physical or technological) is required to register the mental fluctuations of a shopper. What matters is what they finally purchase. (Mind changes are, of course, much cheaper with a Dash Cart.)
Technology can cheapen labor by increasing productivity, but capitalism still needs, and daily depends on, the lifeblood of low wages. Even the magic wand of AI will not make this basic fact disappear. Mikhail Tugan-Baranovsky's economy of machines making machines is still impure. It still has one worker. Labor still has to be there.
Baranovsky writes:
If all workers except one disappear and are replaced by machines, then this one single worker will place the whole enormous mass of machinery in motion and with its assistance produce new machines and the consumption goods of the capitalists. The working class will disappear, which will not in the least disturb the self-expansion process [Verwertungsprozess] of capital.
In our technologically advanced times, Baranovsky's metaphysical one worker is found all over the Global South. They were on the low-wage labor the massive and complex container ship that recently crashed into and brought down Baltimore's Francis Scott Key bridge. They are the dirt-poor laborers suffering Heart of Darkness-like conditions in the human-dwarfing cobalt mines of the Democratic Republic of Congo. (The cobalt is used to make machines like "smartphones, computers and electric vehicles.") Indeed, this is the deepest insight of Alex Rivera's sci-fi masterpiece Sleep Dealer. Though the US's farming, construction, and manufacturing industries are run by robots, these robots are operated by low-wage laborers in Mexico.
