At least five interesting things to start your week

Nvidia's triumph, "greedflation", public housing, Atlanta vs. Los Angeles, and what the UK gets right

I’ve been meaning to do more aggregation — lists of short takes on interesting news and research items that don’t necessarily need a full post. The start of the week seems like a good time to do these. As a compromise, I’m going to make the first part of the list free and the second part for subscribers only.

Also, I’d like to add in a little extra feature! If you’re a subscriber, and you have other interesting links from the past week — news, papers, blog posts, Twitter threads, etc. — that you think the other subscribers would be interested in, let me know in the comments! If I see a couple that I think I should comment on, I’ll pull them up and add them to the list. (This is inspired by Marginal Revolution’s “assorted links” and Matt Yglesias’ “mailbags”.)

Anyway, without further ado, here’s this week’s list.

1. Nvidia’s triumph gives America a hint for how to get industrial policy right

A big piece of news from the corporate world this week was Nvidia’s massive stock surge. After the company announced revenue projections that were much better than forecasts, its stock soared by 24% at the end of the week, giving it a market value of around $1 trillion. Tech now has a new Big Five list — Apple, Microsoft, Google, Amazon, and Nvidia. (MANGA, perhaps?) Jensen Huang, Nvidia’s rock-star founder and CEO, saw his net worth rise to $35 billion. Of course, he celebrated by visiting a night market in his native Taiwan (pictured above).

The reason Nvidia is kicking so much butt, of course, is AI. Nvidia makes chips called GPUs, as well as a proprietary language called CUDA that is used to program these chips. Computing power (or “compute”, as they say) is emerging as a key bottleneck in AI — the new models cost hundreds of millions of dollars to train. GPUs are better for AI than traditional CPUs, because they run lots of mathematical operations in parallel, and AI is highly parallelizable. Thus, Nvidia is incredibly valuable to the emerging AI paradigm.

Nvidia’s victory gives us an important insight into how to use industrial policy to shore up America’s dominance in the semiconductor industry. The obvious lesson is “admit more skilled immigrants”; since Huang came over from Taiwan at the age of 9. But the deeper lesson is about disruption.

Nvidia disrupted the now-struggling Intel, in the classic Clay Christensen sense of disruption — it made a product that at first served a niche market (computer graphics), and which later took over much more lucrative and generalized markets. Intel, with its relentless focus on making incremental improvements to CPUs for servers, missed out on the GPU explosion. It also missed out on other disruptive innovations — TSMC’s foundry model for manufacturing, and cell phone chips.

But Intel itself was the product of successful disruption, back in its day! Intel was initially making memory chips, in a market that was increasingly commodified. So it invented the microprocessor, which at first didn’t have many buyers but eventually became a much bigger and more lucrative market than memory. It was Intel’s disruption of the industry, more than U.S. efforts to support our memory makers, that ensured two decades of American dominance in semiconductors. Nvidia is mainly just a replay of that story — and TSMC, which created the foundry business model from scratch with the generous support of the Taiwanese government, is an opportunity the U.S. missed.

So the big lesson from Nvidia is not that U.S. industrial policy should support Nvidia — it’s that U.S. industrial policy should support the scaling up of startups in the chip industry as a whole. Nvidia itself is highly likely to get disrupted at some point, as is TSMC, and we don’t know what will disrupt them. But we do know that whatever does disrupt them, it ought to be an American company that does it. We need to be focusing on helping a bunch of companies that might be the next Nvidia or the next TSMC, instead of just throwing a bunch of government money at big incumbents.

2. Where the UK is beating the U.S.

I am still a bit annoyed at the Financial Times’ John Burn-Murdoch for claiming last year that the United States, along with the UK, was “a poor society with some very rich people”. That was simply false, as I showed in a lengthy post. The UK sort of deserves the epithet, but the U.S. does not — our middle class is richer than that of almost any developed country.

But Burn-Murdoch is on much more solid ground when he lambasts the U.S. on its health outcomes. Here is an area where the UK, for all its material shabbiness, is outpacing the United States:

Source: John Burn-Murdoch

This is also not due to racial demographics. As Burn-Murdoch shows in a follow-up column this week, Asian and Hispanic Americans have life expectancy that compares to the British, but White and Black Americans are both far behind:

Source: John Burn-Murdoch

This is something that we should be very concerned about. Material goods and services are not the only measure of how well a nation is providing for its populace — public health also matters a lot, even though it doesn’t show up in the GDP numbers. And it’s here that the U.S. is failing its people, while the UK, despite lagging in material welfare, is succeeding. I’ll have more to say in later posts about how we might turn this around (and why our political economy makes it difficult).

3. I disagree with Matt Yglesias on public housing