Re: re: Seems like this might be something AI could usefully be applied to…
The rest of the comment may convince you that I didn't have an LLM write it for me. However, whether I did or not, I used those two phrases to refer to two problems with concerns about novel tracking techniques. Unprovable is a concern because, if you don't try to prove that something works the way you've described, then how do you know that it does? If you accept your theory about how something works when you cannot test it or even demonstrate that something has definitely happened at all, then why should anyone believe it happened rather than making up whatever story they like and believing that without trying to prove it.
The faulty assumptions lead us there and it's something that everyone is vulnerable to. We all have situations where we think something has happened when it has not. Here is an example:
A while ago, I opened Notepad++ on a Windows computer and a window popped up. Well visually it didn't, but I have some software monitoring for invisible windows which noted that it did, and it had a scary name: "Input capture". I've seen that window before, and it most often happens because of a remote desktop connection or a GUI VM connection. What could cause this? I was sure that I hadn't done anything unusual lately, so it clearly couldn't have been me. Maybe I had malware from somewhere. Maybe this was evil Microsoft snooping on me. Maybe someone had infiltrated Notepad++ and given me a poisoned update. Something was being done to me and I was determined to stop it. A full disk scan identified no malware, and a reboot didn't make this go away.
As it turns out, it was me. I had started a WSL VM in an earlier session and opened a file from that session in Notepad++. Then I rebooted, so WSL wasn't running anymore. When I tried to open Notepad++, it tried to open a path that looks like \\wsl.localhost\Ubuntu-22.04\... and that started some searching processes looking for the remote computer so I could get the file I had requested, but they didn't work because the remote computer was on my computer and I hadn't turned it on. I didn't see that because Notepad++, not getting the file it asked for, helpfully didn't open it and showed me the next one in line. That was not in my list of assumptions.
If you see an advertisement for something, there are lots of reasons it could have gotten there and some of them are simple enough that we don't consider them. We may assume that we never visited a page about this topic, even though we actually did and a tracker tracked us from it. This is logical because visiting one website is a really small action, especially if we just glanced over it before abandoning it as not a useful site, so it got about four seconds of our attention. We don't need to remember every little thing like that, so we forget it, and then when something crops up days later, we have forgotten that easy step. The problem with assuming something we haven't proven is that, if it's related to something we can prove, then we can stop it or at least try to. If we always assume that we have stopped all the basic web trackers, therefore any tracking we see must be due to something else, the chances are high that we haven't stopped all web trackers and we will get more done by finding the ones that slipped through.