Operating systems are built on conventions nobody chose deliberately and on maintenance work nobody wants to fund. That is the common thread in three recent items on this beat, and taken together they suggest the costs of both are becoming harder to defer.
The 1970 Problem Is a Design Choice
Engadget's recent piece on the arbitrary dates that appear on non-Windows devices makes a point that is easy to skim past: when a device shows a date in 1970, it is not glitching. That date was chosen during the development process. The detail matters because it is the clearest kind of operating system convention - a decision made early, for reasons that made sense to the people making it, and then carried forward for decades because changing it would break things.
For US consumers, this is mostly invisible until it is not. A date stamp on a photo, a file timestamp, a log entry can all inherit a convention that no product manager today would pick. The lesson generalizes across the operating systems Americans use every day. Much of what feels like a bug is a legacy choice preserved on purpose, because the cost of revisiting it would fall on developers and users who never signed up for the change.
That does not make the convention wrong. It does mean the industry's default answer to awkward foundational decisions is to keep them. Every year that answer holds, the accumulated weight of compatibility grows.
Maintenance Is the Real Bottleneck
ZDNET's report on Linux bug-finding and bug-fixing pulls the same thread in a different direction. Finding Linux bugs is easy now, the piece argues; fixing them is another story. Some maintainers are quitting under the strain of AI-generated bug reports, while others see signs for hope.
The asymmetry is the story. Tools that generate reports have become cheap and abundant. The humans who triage, validate, and repair have not. An operating system is not secured by the volume of issues filed against it; it is secured by the judgment applied to those issues. When the filing side scales faster than the fixing side, the bottleneck moves to the people, and some of them leave.
The American angle here is direct. US companies build enormous commercial estates on open-source operating system components, and US consumers depend on them without ever seeing a maintenance contract. The institutions doing that upkeep are not American corporate departments with headcount budgets; they are distributed maintainer communities. The strain ZDNET describes lands on infrastructure that US firms treat as free.
What the Two Stories Share
Put the Engadget and ZDNET items side by side and the pattern is not about dates or bug trackers. It is about the unglamorous substrate of operating systems - the conventions that persist because change is expensive, and the labor that persists because someone cares enough to do it. Both are forms of accumulated debt, and both are easy to ignore while they hold.
AI changes the arithmetic on the second. It lowers the cost of producing plausible-looking input to a system and leaves the cost of adjudicating that input roughly where it was, or higher. A convention like a 1970 default survives because the people who understand why it exists are still around to defend it. When maintainers exit, that institutional memory goes with them, and the convention becomes harder to change, not easier.
