Longreads + Open Thread

Longreads

• Tyler Cowen interviews Jonathan Haidt about Haidt's new book which argues that smartphone addiction is a pressing issue. Cowen raises some good points: banning teens from using social media requires some pretty intrusive rules about what parents can allow their kids to do, or requires online services to set up some very annoying roadblocks that won't necessarily work. But the pro-social media view also has some problems—it's true that new technologies are disruptive before we develop norms around their use (stop measuring at the right time, and the biggest visible impact of the printing press is the Thirty Years' War), but the way we deal with those technologies often is a shift in laws and norms. There's a fun exchange where Haidt asks a hypothetical about switching to a diet without vitamin C, and just learning to tolerate scurvy; Cowen points out that the Aztecs added nixtamal to their food to deal with vitamin B deficiency; what Haidt doesn't say, but could have, is that this eliminated the deficiency and that his policy proposals would ideally have the same effect. I came away from this post with a more elitist view of social media—Cowen argues that it accelerates young people's networking at a critical time, and that is true in my anecdotal experience. If it's good for the most smart and driven, but not as good for everyone else, it's hard to compute the costs of a ban. So the boring answer is that, like many other decisions, it will ultimately come down to a balance between a) parents trying to do what's best for their specific kids and b) kids doing the usual thing where they put enormous effort into circumventing rules, but maybe learn something along the way.
• John Herrman traces the origins of porn-bot spam on Twitter. It's as dreary as you'd expect: it’s all scam ads for scam websites, a model that can survive the low revenue per click because there are just so many clicks available on a site that doesn't effectively moderate spam. And while it's not accurate to say that this kind of spam is a trivially solvable problem, it's also true that it's not a problem other sites operating at Twitter scale have failed to solve.
• Ben Reinhardt writes in defense of academia. Some of his arguments apply to other institutions, but this one stood out: "Academia is one of the only places where active, cutting edge research labs are incentivized to train novices. A rule of thumb is that the average grad student is a net drag on productivity until their third year. No other institution has developed structures to support an unproductive worker for three years until they become productive." There are some firms that have turned noncompetes and deferred bonus compensation into an art form to get this kind of result, but it is fair to say that in a liquid labor market, the optimal thing for a self-interested person to do is to work at the company that can train them for a few years, then jump to a company that can pay them more because it doesn't have to subsidize so much employee training. Oddly enough, the industry that has done that exceptionally well is the cozy, uncompetitive, communitarian business of asset management, particularly private equity: their personnel model is essentially that banks train analysts for a few years, and then those analysts leave banks to take better-paying jobs at funds, but the funds reciprocate by sending lots of business back to those banks, which then use the deals they get to train the next generation of analysts. But that's an emergent process, not one that was engineered top-down.
• T. Greer in Scholars Stage reviews Wang Huning's America Against America, a book originally published in 1991 that's gotten more attention as Huning has gotten more prominent in China. Travelogs are information-dense because they highlight things that only a visitor would realize are newsworthy surprises, so someone moving from one end to the other on the cheap stuff and precious time or cheap time and precious stuff continuum will notice a lot. But a lot of what they'll notice will be particularly salient to a poorer country, and less action-guiding for a rich country trying to understand itself.
• In the NYT, Patricia Cohen covers the new model for catch-up growth. It's hard to compete on low-cost manufacturing because China did that so effectively, is so big, and has been relatively good at surviving the transition to higher wages by investing more capital rather than outsourcing to the next cheap labor market. So instead of following the East Asian development model, countries have to try the riskier approach of leapfrogging to an export-driven services economy.
• This week's Capital Gains, linked two bullets above (and quite relevant to the bullet above as well) is about how bad service and expensive nannies are symptoms of a richer economy, and are the flipside of being able to replace a broken product without a second thought instead of laboriously repairing it until it's at the end of its life.
• And in the ever-more-popular Riff podcast, we're talking about corporate espionage, outsourcing, Apple's strategy, AI creating jobs, and more. Listen with Substack/Spotify/Apple.

Books

Means of Control: How the Hidden Alliance of Tech and Government Is Creating a New American Surveillance State. A book much less alarmist than its subtitle. The main thrust of the book is that historically, the organizations that were most interested in, and best equipped to, assembled detailed dossiers about people and track their movements were law enforcement and intelligence services. But the online advertising ecosystem and certain aspects of the Internet of Things have, without directly trying to, created massive collections of exactly this sort of data, and governments have learned to buy it and use it.

That's created all sorts of conflicts: adtech companies don't necessarily want their data to be used for intelligence-gathering and law-enforcement, and will sometimes block these users when they see them; consumers don't like the idea that they could be accused of a crime based on metadata, or that random law enforcement officers can snoop on them (one problem mentioned in the book: when local police departments get some fancy new investigation tool, some members invariably use it to spy on partners, exes, and the like). All that discomfort creates an incentive to use the data, but quietly, and to prefer parallel construction to transparency.

This debate will continue, and in a sense it's older than the rise of real-time bidding infrastructure for online ads in the 2000s. For example, it might raise issues if it became apparent that US intelligence operatives were engaging the services of an independent intelligence-gathering organization that dispatched agents around the globe to gather human intelligence without any direct government oversight, and without ever revealing who they gathered information from. But that's just a technically true way to describe a CIA employee reading a copy of the New York Times.

Open Thread

• Drop in any links or comments of interest to Diff readers.
• The usual trend in programming has been to move to increasingly high-level languages, from assembly to C, from C to Perl and then Ruby and Python, and from raw Javascript to frameworks. And we're moving even further than that as some programming tasks get done by using an LLM to compile something into a high-level language. But there are some corners of the world where performance is paramount and the low-level languages still matter. If you work close to the metal today, especially in a non-HFT context, and would be willing to chat on- or off-the-record, we'd be delighted to talk to you.

Diff Jobs

Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:

• A fintech startup that lets investors trade any theme as if there were an ETF for it is looking for a senior backend engineer. (NYC)
• A company building ML-powered tools to accelerate developer productivity is looking for a frontend engineer with an eye for UX. (Washington DC area)
• A company building the new pension of the 21st century and building universal basic capital is looking for a product designer with fintech experience. (NYC)
• A seed-stage startup is using blockchains to enforce commitments and is in need of a fullstack developer with Solidity experience. (Remote)
• A CRM-ingesting startup is on-boarding customers to its LLM-powered sales software, and is in need of a product engineer with a track record of building on their own. (NYC)

Even if you don't see an exact match for your skills and interests right now, we're happy to talk early so we can let you know if a good opportunity comes up.

If you’re at a company that's looking for talent, we should talk! Diff Jobs works with companies across fintech, hard tech, consumer software, enterprise software, and other areas—any company where finding unusually effective people is a top priority.