This is something that freaks me out about modern tech. Failure rates might be low, but a failure could be catastrophic for the person impacted. I remember se

This is something that freaks me out about modern tech. Failure rates might be low, but a failure could be catastrophic for the person impacted. I remember seeing something similar in Ubiquiti's UniFi forum a few years back. Their WAPs and switches connect to a management portal you can self host and, after treating both http and https interchangeably forever, they decide https should be actual https. Before that, you could type https:// https://..., but the communication would not use TLS. They had customers with misconfigurations because they were consistently typing their inform (aka server) URL as https:// https:// instead of http:// http:// during device configuration. As soon those devices were updated, the new firmware would expect a proper TLS connection for https:// https:// URLs, so it would fail. I remember one person in the forum thread saying they bulk updated 600 devices spread across 300 sites and they all disappeared. That person made mistakes, but the configuration they were using was working, so why would they expect that to change? I tried to make the case that it should be considered a serious bug and part of the response was: > We discussed this internally and decided to move forward because 1) we didn't feel that there would be many affected That was a real eye opener for me and now I operate under the assumption that most tech companies won't have an issue completely ruining your life as long as it's only you or a small enough group of users that it won't have an impact on their bottom line.