Everything is broken and that's great

A friend of mine recently left his bigtech software engineering job and started his own thing. He’s been working on it for a few months now. The previous company was full of Xooglers so they were very much in the thrall of Google’s infrastructure - where open-source versions didn’t exist they built their own.

Now in the Real World™ and without a huge developer experience team he's having to use more off-the-shelf software. He complained that GitHub Actions has basic issues with caching and bandwidth - caches not being correctly invalidated leading to things having to be rebuilt with every test run, or slow cache loads causing slow builds (and therefore more Actions minutes and thus a bigger bill). With a few hours' work he was able to improve on their offering.

(Don't believe me? This tiny outfit publishes set of tools for running builds on GitHub Actions much faster, with much faster throughput, and at a fraction of the cost. We relied on it heavily at my last company and it was a hilariously good deal.)

We should be enormously grateful that this is the state of the world. This is the basic premise of startups: that there is low-hanging fruit waiting for the picking basically everywhere you look. The basis thesis of startups is totally implausible - that a tiny team can take on a goliath corporation with millions of dollars allocated to R&D and build something better. For this to be even remotely possible, there have to be things just waiting to be fixed, or invented.

And it turns out this is the case in basically every domain. Almost every piece of software you interact with could be far better. Not just crummy UX or buggy login flows or whatever else, but 10x improvements that come from seeing things with a fresh pair of eyes and without the bureaucratic cruft that seems to make basically every company impossibly sclerotic with time.

Brokenness is validation that your market opportunity is real and substantial. Most successful startups don't create new markets – when seeking funding, most founders argue that their thing creates some brand new market that inexplicably no-one has thought of before. But perhaps founders should be aiming to enter existing markets and fix their rampant brokenness.

Small teams can move faster and focus more intensely on pain points that larger companies either can't fix or don't notice. The abundance of eminently fixable problems means that basically endless startup potential exists.

This is part of the reason why I don't think software engineers need to worry about their profession becoming obsolete (at least for now). AI lets smaller teams build things much, much more quickly, smashing out new prototypes and slamming them against the world with impossible velocity. The world's demand for software seems nowhere close to being saturated. There's money on the ground for those willing to look for it.