Legacy tech migration is harder and later than it should be
EVE Online's migration from Python 2 to Python 3 landed on the front page and the thread immediately went to Goldman Sachs and JP Morgan maintaining Python 2.7 as a core language until just a few years ago. The point isn't that EVE Online is special. It's that large codebases with real stakes stay on deprecated runtimes far longer than anyone admits publicly, and the migration story is treated as remarkable when it should be routine.
The FreeCORE TrueNAS fork is a related signal: a subset of users cared enough about staying on FreeBSD that they forked the project rather than follow iXsystems to Linux. That's a lot of effort to avoid a migration, and it tells you something about how painful these transitions feel even when they're technically tractable.
RISC-V getting Tier 3 CPython support is the optimistic version of the same story: new platforms take years to reach the point where runtimes support them officially, and Tier 3 still means it can break without blocking anything.
So what?
If you're inheriting or acquiring a codebase, the Python 2 situation is a useful reminder that technical debt in runtimes is not self-liquidating. Someone has to decide to pay it, and that decision keeps getting deferred. Building migration plans into your roadmap from the start is cheaper than the EVE Online version.