This particular effect has nothing to do with whether or not anyone is looking at an object. It's the methods of making it visible that cause the effect.
I think that a big part of the problem is trying to make the new more layman friendly removing all the math. Quantum mechanics is unintuitive but the math is very clear, so "anyone" can get the same correct result for an experiment. When you remove the math, you keep only the unintuitive part.
Another part of the problem is linkbait, making slightly wrong interpretations make more interesting articles. Sometimes there are a few articles from different news sites about some subject, and it's very common that the more misleading gets more attention here.
Sadly, my rule of thumb is https://xkcd.com/1240/
For an interesting explanation of what is happening here, I like the stack of slightly rotated polarizers. (See this comment https://news.ycombinator.com/item?id=10437289 ) The math is equivalent, but the result is not linkbaity. You can even try it at home! If you think enough about it, it's completely weird, but the calculation are pretty straightforward.
Generally a decent layman's explanation of the experiment, but they had to conflate quantum mechanics with magic by saying "the atom knew it was being watched..."
It's very hard to handle these misconceptions once they're already so widespread.
Calling printf can (will) change what's beyond the stack of the calling function, so if the pointer in question pointed outside the active stack, printf may have put a safer value there by coincidence.
methods of disassociating mutable objects before printing will save you from some heisenbugs
a la javascript
console.log(JSON.parse(JSON.stringify(obj))https://www.youtube.com/watch?v=sQfSm6o-KlQ
(Assuming it's correct - I'm not physicist)
For example, in the quantum eraser video referred to by the video you linked, it's said that there's an interference pattern for B when you don't measure the entangled associated photon A (and that there is one when you do measure A). That's not the case. If it was the case, we'd have a handy-dandy FTL communication mechanism.
What actually happens is that there's never a visible interference pattern for B in this experiment. Instead, you use the measurements of A to filter or split the measurements of B into two groups (e.g. Bs when A up, Bs when A down). Within each group of measurements, you'll find an interference pattern. The two interference patterns will complement each other so the sum is an apparent lack-of-interference; you need the As to do the separation.
I personally think the the quantum eraser experiment has a terrible misleading name. It's more of a "use entangled measurements to find the interference" experiment.