On Secret Weapons and Magic Bullets

There's Never One Weird Trick

In partnership with

Know someone who might like Capital Gains? Use the referral program to gain access to my database of book reviews (1), an invite to the Capital Gains Discord (2), stickers (10), and a mug (25). Scroll to the bottom of the email version of this edition or subscribe to get your referral link!

When a successful company or individual gets profiled, the subtext is: “what could the audience copy to be just as successful as the person being profiled?” This is unavoidable; there isn't a lot of demand for profiles of people who achieved 70th percentile outcomes in their chosen field, unless that outcome is 99.9th percentile when we account for their circumstances.

And successes demand an explanation from the people responsible for them, too! Nobody wants to read a profile of a successful investor who claims that it all came down to luck, or from a company that attributes their outsized returns to a fortunate mistake made by a competitor that they exploited well.

One instance of this is big bets on unusual technologies. There are many examples: Meta would be a much smaller company if they hadn't decided to win in mobile or die trying; TSM is a bigger business thanks to their embrace of EUV (or are they a big business because Intel didn't?); Nubank has been able to recruit more technically skilled engineers because it's a Clojure shop, even though it's also a bank.

But these bets work better as narrative hooks for a more boring and harder-to-emulate story of success. How does a company become a big user of some obscure language? It's probably not because a nontechnical founder read a really compelling paper on compiler design! No, the usual story is either that the technical cofounder made an arbitrary upfront decision, or that at some point in the company's history, somebody got frustrated with the pace of development or the quantity of bugs and said "You know what? If we were using Erlang, I could build this in a weekend."

The company will not later be lauded for its bold decision to embrace Erlang, the official language of phones not dropping calls and planes not dropping out of the sky, if this person does not, in fact, produce a lot of working code over the course of that weekend. There are plenty of companies that make unique technical choices and just flail around for a while before quietly shutting down, especially when the reason for the startup was to find some excuse to use a fun tool.

So you need a sort of boldness/outcome fit: the kind of company that would do this is different from the kind of company that wouldn't, and the kind of company that would get it right is even more different from that. Making an unusual technical choice is a bet that:

  1. The language or tool itself is genuinely better than more popular options, i.e. the same developer, once they're up to speed, is more efficient.

  2. There's a selection effect where people who are opinionated about languages are better employees than people who aren’t.1

  3. These choices are a retention tool, in two ways: there are fewer places where people can go to use the language they like, and the skills they've gained on the job don't apply well elsewhere.

  4. They won't fully tap out the pool of people who want to, or are willing to, use the language of choice. Alternatively, they expect to be so successful that they can singlehandedly support the language themselves, contribute to open-source libraries, and otherwise keep it alive. This last one is one reason you can point to big successes that attribute their success to choosing an unusual language: it’s a high-variance bet, and if they’d chosen the wrong language or hadn’t built a big enough business to independently support that language, they could end up left behind, spending all of their time building libraries and integrations that other language users get because the open source community builds them.

Like any bet, these are all uncertain, but the fact that they're uncertain means that such choices are a strong signal. Every organization has a finite number of Weirdness Points, which are earned at roughly the log of their valuation. If a company does everything differently—no org chart, an unusual structure for employee equity, a business model nobody else believes in, total salary transparency, etc.—they'll burn all of their credibility. So the best companies tend to make a short number of distinct choices, and otherwise run everything either a) in a completely normal way, or b) in the most normal way they can considering their other decisions.

So the ruthlessness of the weak form of the efficient market hypothesis strikes again: the "silver bullets" that universally work are the ones that every company already does. Which means that there's an infinite list of pointless ones—you're guaranteed to be a more successful company if you don't commit crimes, and if you commit to paying people more than 10% but less than 1,000% of what people earn for similar jobs, etc. And the actual weird choices that do correlate with success do so because they're part of a framework that actually made that success more likely. It’s probably true that a company that can go all-in on writing everything in one language that just barely misses the cutoff for Wikipedia's list of esoteric programming languages has better odds than one that doesn't. However, 1) the odds are higher if they could do that while spending their weirdness points elsewhere, and 2) whatever combination of trust, talent, and openness that made that choice feasible was probably the underlying reason the company was going to win.

Disclosure: Long META.

Boost Marketing Results with Anyword’s Performance Prediction

Meet Anyword. The AI that knows your brand, your audience, and what content resonates.

Trusted by 1M+ companies, Anyword’s AI connects your marketing channels to analyze past performance and scale high-performing, on-brand content.

Read More in The Diff

The Diff spends a lot of time covering what makes different companies unique. 

1. Importantly, this is a bet that can go in both directions, especially because of the dynamic from this classic joke: perhaps your team is completely on board with doing everything in Haskell, and then loses weeks of productivity each year to debates over which editor to use.

Share Capital Gains

Subscribed readers can participate in our referral program! If you're not already subscribed, click the button below and we'll email you your link; if you are already subscribed, you can find your referral link in the email version of this edition.

Join the discussion!

Join the conversation

or to participate.