So, this is one of those ideas that's been such a long-term preoccupation of mine that I just assume I've written about it before, probably multiple times, but: as we make more and more engaging signalling and entertainment, the optimization process can be done without regard to physical constraints. You can make a comfier couch, if you have better materials, but materials don't spring up ex-nihilo – words, images, and video, essentially do, as modern editing techniques become extremely sophisticated. The couch has to be possible – a video game's story doesn't.
Don't get me wrong, we ought to worry we make chairs so comfortable that people sit in chairs so much it harms them. Chairs are within an optimization engine (known as 'people being useful to each other') that requires wisdom when making decisions. Perhaps the even more comfortable chair isn't good for you – I don't know. But without wisdom, you'll indulge a simply described desire too much, or short-term preferences over harder to articulate long-term preferences, and be worse off.
But when you combine that general society-optimization process with, essentially, unbounded capability, we ought to expect, e.g. the content on our screens to be superstimulus, hyper-compelling, dangerous, and wise to use very carefully. I think that point has arrived for phones. At least people can (and do) use them on the toilet so they don't die like the people who passed while playing World of Warcraft.
When we read Eliezer Yudkowsky saying 13 years ago that this is a potential long-term threat to humanity, we ought to ask ourselves if we can make meaningful statements about the nature of the threat. It's also worth echoing the comment that the threat isn't violent video games, it's children prefering Megaman to playing outdoors. Drops in violence would be an early indication that perhaps the superstimulus was increasingly filling the needs of intense stimulus-seeking aggression.
They also note that we need to give people better and better tools for self-control, or rather, explicitly designing their own experience. Modern cell phones have made this challenging, given you essentially need them and the tools for control are weak. I also think it's relatively clear that books are much better today than they were 40 years ago, fiction and non-fiction (although non-fiction is a bit harder to find excellent examples of, but most of that has to be terrible books sucking a lot of oxygen out of the room). Things are progressing, but it's not clear to me they are accelerating – but the size of the problem does seem to be going up, and not down, so it's probably worthwhile to discuss the future.
I'm most worried about VR as a whole-sale experience, as well as in-home AR with mechanical peripherals. It isn't passive, like video, and doesn't have the casual relationship with time books do (the best part is jumping around on a page, piecing ideas together across a chapter, etc). We're also seeing more ephemeral and ever-changing game ecosystems (like Fortnite) that might plausibly induce an upward sloping demand curve with percent-of-time-invested. That seems to be the most dangerous thing. I'd be shocked if neither a coherent control mechanism nor an upward-sloping-demand game wasn't made by 2030.