> You’re driving somewhere important…maybe running a little bit late. A few minutes into the drive, your car pulls over to the side of the road and asks: “How are you enjoying your drive so far?” (...) “Did you know I have a new feature? Tap here to learn more.”

> There was a time when software shipped on physical media: floppy disks, CD-ROMs. (...) If you shipped a serious bug, you lived with it until the next release. (...) Then the internet arrived. (...) Bugs could be fixed. Security holes could be closed. Documentation got easier. Support got easier.

> Once software could reliably connect to the internet, it no longer just received instructions. It could talk back to the company. (...) Instead of arguing in a meeting about which features mattered most, you could look at usage data. (...) Usage data stopped being a tool for improving software and became a tool for optimizing behavior.

> The product stops being a finished artifact and starts behaving like a laboratory. (...) Every interaction becomes a hypothesis. (...) Iit’s very hard not to optimize for what moves fastest, even if it moves the wrong thing. (...) You add a little callout. Then a tooltip. Then an onboarding tour. Then a “What’s New” screen. Then a little survey.

> Not every prompt is evil, and not every notification is marketing. Some categories are basically made of alerts: messaging, security, banking, calendars, delivery tracking. (...) Once a company builds the machinery to do it, that machinery becomes cheap to reuse, and the incentives gradually pull it away (...) toward “move the metric.”

> Great tools get out of the way so the user can accomplish their goal. (...) You open them, do the thing you came to do, and close them again without ever feeling managed, marketed to, or delayed. (...) If you want to survey me, fine. Ask at a sensible moment and accept “no” as a real answer. Most importantly, if I turn something off, it should stay off!

> Each step made sense in isolation, and each step could be defended. Together, they reshaped the priorities of an entire industry. Once software became measurable in this way, it became optimizable in this way. And optimization has a way of eating everything else.

https://blog.mikeswanson.com/backseat-software/