Lindy Effect

This is such a useful concept. I work in an organization that owns dozens to hundreds of different network services, data pipelines, and other applications. I haven't collected the data, but it seems to me that their useful lifespans follow Lindy's Law. For example, I'm currently seeing a crop of deprecations of systems that were created in the last five years, while systems that are a decade or older show no signs of going anywhere.

I've found it a good career niche to focus on these older systems. Most people are attracted by the new and shiny (LLMs, anyone?) but the older services have proven themselves to have durable functionality. Quite often there are interesting opportunities to improve, augment, or even replace these systems if you only look for it. When done well you can provide a ton of long-lasting value.