Longreads

• A good look at construction startup Katerra, from a former employee. The thesis is that construction is unnecessarily expensive because so many components are one-off rather than modular; there are diseconomies of scale built right in. This is partly driven by the other, bigger problem in housing costs: it's legally hard to build in parts of the US where people want to live. This imposes a high fixed cost that makes other costs less salient. Katerra was not unaware of this, and there's a near possible world where a successful version of the company was able to lobby for more flexible land use regulations. It's a great story about the challenges of scaling. Great companies go through a period where they grow right at the limit they can handle, and Katerra tried to grow a little faster than that.

• A good contrast with the above is this reflection on scaling YouTube. The key theme is to have forward-looking focus and to be thoughtful about priorities, and to add some deliberate slack to the schedule. A lot of these practices coevolve with the team the company recruits, so they don't necessarily apply elsewhere, but they do show what a good process looks like when it's mature.

• Apple's new email privacy settings make life more difficult for some publishers, including newsletter writers. If nothing else, this is evidence that Apple isn't solely committed to making Facebook's life difficult—given a constant level of content creation, bad news for email distribution is good news for social media. This will be much worse for marketing emails than for newsletters as such, and like many other privacy enhancements it ends up rewarding the largest participants. A large enough newsletter can estimate audience activity from link clickthroughs, but at a smaller sample size there's just not enough data.

• Thom Lambert of Truth on the Market compares the Grail/Illumina merger to a Theranos-that-works, and notes that it's being delayed. This is a complicated story involving the interaction of two different regulatory regimes: healthcare rules make it challenging for small companies to be full-stack rather than leaning on partners to monetize. But antitrust regulators are naturally concerned when a dominant company uses mergers to vertically integrate.

• "Startup valuations seemingly live in the corners of the valuation room that the DCF Roomba cannot reach." A good post from VC Sajith Pai on how VCs think about valuation. At some level, businesses valuations are ultimately constrained by the net present value of the cash flows that business can produce. The DCF valuation is always somewhere in the distance, even if it's not visible yet. So startup valuation is moving from one landmark to the next, even when the destination isn't visible yet.

Books

• Destined for War advances the thesis that competition between the US and China matches a historical pattern that usually, but not always, leads to war. The thesis is actually fleshed out quite responsibly in the book, with careful attention paid to the cases where wars didn't happen. The world is generally less violent right now than it was during most of the other case studies. But many of those case studies were preceded by periods where war was either rare or one-sided (or, really, 1.5-sided: politically, it's very domestically unpopular for a big country to appear to be pushed around, but wars that aren't as easily won as they look like quickly get unpopular).

Open Thread

• Feel free to post any links you think Diff readers would enjoy.

• Next week I'll be taking a look at some upcoming IPOs (Didi, Clear). They're very different parts of the travel industry, but a year of mass relocation is a year in which some long-term travel trends permanently stopped and others have started. Are there any interesting views on the future of travel? It's been interesting to see small datapoints about the return of business travel—blow-by-blow accounts of merger negotiations still have lots of in-person meetings, for example.