I would be surprised if Andreessen’s highly educated readers really believed in the pooled labor fallacy, but anyway he goes ahead and breaks it down, introducing – as if it were new to him. to his readers – the concept of productivity growth. He argues that as technology makes companies more efficient, they pass the savings on to their customers in the form of lower prices, leaving people with more money to buy more things, do increase demand, increase production, in a nice self-sustaining way. good growth cycle. Better still, because technology makes workers more productive, employers pay them more, so they have even more to spend, so growth doubles.
There is a lot wrong with this argument. As companies become more efficient, they will not pass on savings to customers unless they are forced by competition or regulation. Competition and regulation are weak in many places and many industries, especially where companies are growing larger and more dominant – think big box stores in towns where department stores are located. Local is closed. (And it’s not like Andreessen isn’t aware of this. His article “It’s Time to Build” against “market-based competitive impediments” such as oligopolies and regulatory capture. )
Furthermore, large companies are more likely than smaller companies to have the technical resources to implement AI and see meaningful benefits from doing so—after all, AI is most useful when it comes to AI. have a large amount of data to process. So AI can even reduce the competition and enrich the owners of the companies that use it without offering discounts to their customers.
Then while technology can make companies more effective, it’s only sometimes made for individuals Worker higher productivity (also called marginal productivity). Other times, it only allows companies to automate part of the work and employ fewer people. The Book of Daron Acemoglu and Simon Johnson Power and ProgressA lengthy but invaluable guide to understanding exactly how technology has influenced work historically, calls this “relative automation.”
For example, take the supermarket self-checkout kiosk. These do not make the rest of the checkout staff more productive, nor do they help the supermarket attract more shoppers or sell more products. They just allowed it to let go of some employees. Many technological advances Maybe improve marginal productivity, but—the book argues—do they DO depends on how companies choose to implement them. Some uses improve workers’ abilities; other things, like trivial automation, just improve the overall bottom line. And a company usually only chooses the first if their workers or the law force them to do so. (Listen Acemoglu talks about this with me on our podcast Have a good future.)
The real concern about AI and jobs, which Andreessen completely ignores, is that while many will lose their jobs quickly, new types of jobs—in new industries and markets created by AI—will It takes longer to show up, and for many workers, re-skilling will be difficult or out of reach. And the same has happened with every major tech upheaval to date.
When the rich get richer
Another thing Andreessen wants you to believe is that AI will not lead to “severe inequality.” Again, this is something of a straw—inequality doesn’t have to be crippled to get worse than it is today. Oddly enough, Andreessen refutes his own argument here. He said that technology does not lead to inequality because its inventors have an incentive to make technology accessible to as many people as possible. As a “classic example,” he cites Elon Musk’s plan to transform Teslas from a luxury brand into a mass-market car — which, he says, has turned Musk into a “human richest man in the world”.
However, as Musk became the richest man in the world by bringing Tesla to the masses, and many other technologies became mainstream, the past 30 years have seen slow but steady growth on income inequality in the US. Somehow, this doesn’t seem like an argument against technology-induced inequality.
Good furniture
Now we come to the logic of Andreessen’s work. Andreessen is right to reject the notion that super-intelligent AI will destroy humanity. He identifies this as just the latest iteration of a long-lived cultural meme about frantic human creations (Prometheus, golem, Frankenstein) and points out that the idea that AI might even be possible decision killing us all is a “classification error”—it assumes the AI has a mind of its own. Instead, he said, AI “is math-code-computer, built by humans, owned by humans, used by humans, controlled by humans.”
This is absolutely true, a welcome antidote to the doomsday warnings of people like Eliezer Yudkowsky—and in stark contrast to Andreessen’s aforementioned claim that giving everyone an “AI coach” will automatically make the world a better place. Like I said: If people build, own, use, and control AI, they’ll do with it exactly what they want to do, and that could include deep-frying the planet.
