Welcome back to Techne! This week I’ve had FACS, a Chicago-based band, on repeat. I’m constantly searching for new rock bands and “Take Me to Your Heart” is one of those tracks that immediately caught my attention. Got a recommendation? Let me know in the comments!

Notes and Quotes

• The Ukraine aid package, recently signed into law by President Biden, includes language from the TikTok divestiture bill and a bill that limits sharing data to Chinese companies, both of which passed the House last month. 
• The Federal Trade Commission approved a rule that effectively bans all non-compete clauses. These contract provisions stop workers from working for competitors or starting a competing business after they leave a job. In response, economist Brian Albrecht explored the limits to the research the FTC cited.
• Meta released Llama 3, an open-source large language model that beats OpenAI’s ChatGPT 4. Dwarkesh Patel’s conversation with Meta’s Mark Zuckerberg offers a lot of context to the release. Apparently, images are produced “so quickly that it actually generates it as you’re typing and updates it in real time,” according to Zuck. 
• The FDA has authorized Sepsis ImmunoScore, a sepsis detection tool that is powered by AI. It’s the first time an AI-powered tool has been approved by the FDA.
• The Future of Humanity Institute, based at Oxford University, closed its doors last week. Led by philosopher Nick Bostrom, the institute popularized the notion of X-risk or extinction risk posed by AI. The closure is widely seen as a setback for the movement Bostrom helped to promote.
• Paul Christiano, a former OpenAI researcher who pioneered a key technique in AI, has been appointed as head of the U.S. AI Safety Institute. Christiano once predicted that “there’s a 50 percent chance AI development could end in ‘doom,’” a term used to describe the complete extinction of humanity. His appointment has some within the government worried that he “could compromise the institute’s objectivity and integrity” and encourage “non-scientific thinking.”
• Amazon has ended drone deliveries in Lockeford, California, one of just two of the company’s U.S. drone delivery sites. I have been bearish on using drones for deliveries, so the announcement seems to confirm my suspicion that the business case is still not there. 
• Josh Dzieza wonderfully captures the hidden world of undersea cables in this long-form piece for The Verge. There are more than 800,000 miles of cable currently laid down, and they are serviced by a couple dozen boats. Just fascinating throughout. 
• I previously wrote about the FCC’s net neutrality rules, but I didn’t mention much about the arguments over investment. New research by economist George Ford finds that “the preponderance of the evidence is against Title II regulation on investment grounds.” In other words, we should expect that the new rules will depress investment in broadband. 
• The consulting firm Multistate has been tracking 685 different bills regulating AI across the country, organized by state.
• The House of Representatives released a report on its efforts to embed AI into legislative policymaking. It’s the first product from the Bipartisan Task Force on Artificial Intelligence and likely not the last, so I’ll be watching to see what they do.  

‘Technofeudalism’ Ain’t It

About a week ago, I was out at a bar celebrating a friend’s admission into law school when a friend and I struck up a conversation with a stranger. After some back and forth over 30 minutes or so, he told us that he was reading Yanis Varoufakis’ new book, Technofeudalism: What Killed Capitalism, likely intrigued by the recent Wired interview with the author. I’m sure I came off as rude, but as soon as he mentioned the title I reflexively said, “Oof, I would be careful with that one.”

Truth be told, I had read only excerpts, so I tore through the book over the last couple of days and it was … largely what I expected. 

While interesting in parts, I found Technofeudalism deeply unsatisfying as a piece of tech analysis. For over a quarter-century, social scientists across disciplines have explored how digital platforms like Facebook and Apple’s App Store price their goods, how users feel about their online experiences, and how these dynamics foster innovation. None of that work, which netted at least one Nobel Prize and has been cited in Supreme Court cases, is referenced or seriously discussed. 

To its credit, the book is pretty sweeping. It includes “a child’s introduction to historical materialism,” commentary on why “Keynes wanted to stop us thinking of money as a thing,” a dive into Thomas More’s Utopia, diatribes against bitcoin, the death of liberalism, and countless other bits. But the driving question of Technofeudalism comes from an interaction that Varoufakis had with his late father, who wondered, “Now that computers speak to each other, will this network make capitalism impossible to overthrow? Or might it finally reveal its Achilles heel?”

Varoufakis’ answer comes in the form of a new theory about tech platforms. He writes:

If we do pay attention, it is not hard to see that capital’s mutation into what I call cloud capital has demolished capitalism’s two pillars: markets and profits. Of course, markets and profits remain ubiquitous – indeed, markets and profits were ubiquitous under feudalism too – they just aren’t running the show any more. What has happened over the last two decades is that profit and markets have been evicted from the epicentre of our economic and social system, pushed out to its margins, and replaced. With what? Markets, the medium of capitalism, have been replaced by digital trading platforms which look like, but are not, markets, and are better understood as fiefdoms. And profit, the engine of capitalism, has been replaced with its feudal predecessor: rent.

Like a puppy that hears an odd sound, my head pivoted and my focus sharpened when I read this passage. The digital platforms that Varoufakis singles out—including Apple, Meta, and Alphabet—are among the most profitable companies ever. It seems like a big stretch to say that profit “has been replaced with its feudal predecessor: rent.” Later in the book, Varoufakis effectively admits that the term “rent” refers to the earnings that come from land, labor, and capital—so rent has never really gone away. In reality, he wants to connect the experience of being on a digital platform with that of feudal serfs who were tied to the land and lived a subsistence life. The problem? I can close my Facebook account, but serfs couldn’t walk away. This mashup of ideas ultimately makes Technofeudalism not just a confused book but also an incredibly limiting one. 

Besides, there is wide agreement in economics, sociology, information theory, and even Supreme Court precedent that these digital platforms are markets. They are a unique kind of market, in fact, that joins two or more groups together for mutual benefit. Along with the wider tech community, economists refer to them as multi-sided markets. 

Social media sites, search engines, and app stores are all multi-sided markets, along with malls, credit cards, and gaming consoles. These markets have been around for centuries, but with the advent of modern communication systems and the internet, the cost to connect dropped, allowing platforms to play the role of matchmaker. The difficulty has always been in finding mutually advantageous pricing and investment strategies that keep everyone on the platform. Think about Facebook: Without good content, users leave, and when users leave, advertisers find other outlets to place their ads. It is this relationship between the two parties that drives user prices toward zero.  

Still, this tendency for social media sites and search engines to be free causes confusion about what’s actually happening. As I have written before,

The term free is deceptive. In one context, free means that a good or service has no explicit price. In this sense, free means that there is zero price at the point of use, such that consumption does not depend on the ability to pay. On the other hand, free also suggests that a thing or service has no cost. But every choice comes with a cost, or more precisely, an opportunity cost. In the classic definition offered by economist James Buchanan, opportunity cost is the anticipated value of  “that which might be” if the choice were made differently. 

This is why I have railed against that old adage in tech: If you aren’t paying, then you’re the product. You are never not paying, though. You are paying with your time. You are paying with your attention. Social media platforms, though free of charge, charge a hidden expense: the opportunity cost of time. Every hour on Facebook is an hour not spent hiking or playing basketball or hanging with friends. And the last time that I added it up, someone using social media for 40 minutes a day implicitly values the site by over $5,600 for a year. For a deep dive into this, check out my piece titled “The attention economy: a history of the term, its economics, its value, and how it is changing politics.”

Reframing our online experiences this way, as part of a multi-sided market, leads us to far more interesting questions than what Varoufakis explores in Technofeudalism. What we should be thinking about—indeed what we are arguing over—is whether or not these platforms are worth the attention that we give them. For kids especially, I’m not sure the easy access to content is a good thing. They should be learning and exploring the world, not spending time in front of a screen swiping. Still, by firmly planting analysis in multi-sided markets, Varoufakis could have better explained why users migrate away from the big platforms, the subtle influence of content moderation choices, and how these tech giants are both allowing new kinds of speech and disallowing other kinds of speech—all of which he tries to tackle in the book. 

Of course, this isn’t the book that Varoufakis wrote. So who out there wants to give me a fat book contract to write it?   

Until next week, 

🚀 Will

Research and Reports

• Right now I am reading a newly released open-access book titled, “Moonshots and the New Industrial Policy.” Lots of money has been spent since COVID on infrastructure, solar panels, and semiconductors, so research of this kind always intrigues me. I’ll report back if I find something that stands out. 
• The economist Daron Acemoglu released a new paper arguing that generative AI systems, like ChatGPT, are likely to have just “a 0.71% increase in total factor productivity over 10 years.” I’m skeptical of the methods, which Tyler Cowen laid out in an extended blog post. But Acemoglu is right to question the orthodoxy that these systems will be a huge game changer.