Homebrew hit 6.0.0 today. 17 years since Max Howell started it as a side project. No venture rounds. No pivot to AI. No rebranding. Just a package manager that kept getting maintained through three operating system generations and five Mac chip architectures.
That kind of longevity is statistically weird in software. Most projects die within two years or get acqui-hired into irrelevance. The ones that survive do it by being boring enough to never need reinventing and useful enough to never be replaced.
Homebrew's longevity is interesting on its own, but what's more interesting is the gap between the attention it gets and the attention launch-focused software gets. When a new framework ships, it gets 500+ HN points, blog posts, conference talks. When a 17-year-old package manager quietly ships a new version with better caching and a few bugfixes, it gets a brief mention and everyone moves on.
This attention gap has real economic consequences. The invisible work of maintenance is structurally under-rewarded relative to the visible work of creation. Venture capital funds launches, not maintenance. Stock prices move on product announcements, not on dependency updates. The entire incentive structure of the software industry is tilted toward the moment code is born, not the decades it might live.
The irony: the software we actually depend on, package managers, operating systems, encryption libraries, programming languages, is disproportionately old. The stuff that lasts is the stuff that got maintained. The 17-year package manager is more critical infrastructure than any 6-month-old SaaS product.
AI coding assistants generating millions of lines of code are making this problem worse. Every line of generated code is a future maintenance liability. The person who generates it won't be the person who maintains it. The launch gets the attention; the maintenance is someone else's problem.
The market doesn't price maintenance correctly because maintenance is invisible until it stops happening. When a package goes unmaintained, that's when everyone notices. But by then the exploit is already in production.
Homebrew 6.0.0 is news because a project survived long enough to reach a round number. That shouldn't be remarkable. But in software, endurance is the rarest quality of all.