A build server that has to be restarted every few days usually has a suite leaking a handful of browser processes per run. After a week there are hundreds, the machine has no memory left, and every job on it is slow.
That happens without anyone writing obviously wrong code. driver.quit() sits at the end of each test, and tests that fail never reach it.
At senior level, driver lifecycle is a resource management problem, and it has three parts: never leak, isolate correctly, and pay the startup cost as few times as you can.