Real investigations are hard, and it's just easier to go after the careless or ignorant, while the actually dangerous and capable evade getting caught.
So, yes, you're right.
No, because content like CSAM will immediately switch channels and you cannot really ban encryption.
The exception is human trafficking and it isn't a crime done by common criminals or lone offenders, it is done by organized criminal gangs in a large scale that can be targeted individually.
I have no obligation to make any suggestion for not wanting my chats surveilled for that matter.
We increased surveillance for decades. And none of this surveillance had any measurable impact on safety. It is just a short sighted proposal and nothing more. We know where and when kids get abused. The abused is almost always in their vicinity. Internet chat surveillance does not help at all.
Not only did the treaty get rejected by some states that were allowed to vote on it, the fact that it is too far removed from citizens and their inadequate participation. In fact I would argue the treaty made it even more undemocratic. The influence of Kutcher compared to European individuals is not the only empirical evidence for that.
The parliamentary control is minimal and there is little overarching political discourse beyond the borders of the nation state. Language barriers are strong and that won't change in the foreseeable future. Until then the commission will run the circus.
It probably depends on the user base of the platform, but I believe it could be one of the worst jobs in the world. If you like the community and content, great, but the usual employed moderators on large social media platforms probably need regular therapy to distance themselves from all the crap they have to read every day. There is a huge difference if you have to read it because it is your job and when you do it on your own because you are hungry for something tasteless and when you can just leave a spicy reply.
Perhaps they believe the product would be cheaper with abdicating the feature?
Or, perhaps, believe it would get updates? Software in cars tends to grow stale quickly, which is an annoyance if you plan to keep for more than a few years. My eight year old car, for example, has features (had features) that don't work properly anymore because the software didn't anticipate how the rest of the world would change around it and there isn't an update provided to restore it to the original working condition, let alone adding new features that would be nice to have.
And car manufacturers are still very much in a mindset that you make a certain model-year, release it, and then forget about it while you work on the next ones. No upgrade path on software whatsoever.
I've driven in a Tesla with FSD, and it's underwhelming. There's no guarantee that in five years, it will perform significantly better - maybe even require additional hardware or a newer car/platform.
In previous versions "Autopilot" as a name only implied self-driving to some customers, while it wasn't. Now they've gone a step further and branded advanced driver assist as FSD, which really isn't all that self driving at all.
But based on our top-of-line 2019 RAV4 the computerization of the car is way more brittle and pointless than the traditional engineering. I expect to have a collection of dead stuff long before the car dies.
If management cannot define a sensible output at the end of a work shift, they seem to be overtaxed with their responsibility.