This is a problem I've been toying with for more than six years, ever since I realized that tech's purpose is to gain control of as much attention span across the planet as possible. It can't help it.
Since I teach how to organize project/product/org information, I played a little mental game: what is the minimum amount of tech I would need and still stay connected to the rest of the planet?
My answer? A piece of hardware that displays plaintext, in e-ink format (to prevent the necessary communications around "upgrade your device now!"). Plain lists of stuff I consume that I can manipulate with my fingers (to prevent keyboards, audio sensors, kinnect, etc from getting their foot in the door). No visible O/S or apps. (No updates, patches, app-store chicanery etc.). Important: no way to install anything or to consume any other content besides what's on the app. No links, no follow-ups. Just what I have predetermined I want to see.
UI? A plain, unadorned list and for each list item some more plaintext.
But what about conversations? Saving stories? Doing research? Well, most of that is social-media addictive nonsense (you really don't need to Tweet "OMG!", but some it is required. For that we have buttons and a microphone. The microphone (and WiFi) have real, wired switches to turn them off and on that can't be disabled by software. (More telemetry problems here).
Four buttons. That's it. The device should try to sort your list and associated text, so you need a way of saying "I like this kind of thing" and "I don't like this kind of thing" so it can learn. You need a way of saving something for later to download, research, study, reply, and so on.
The fourth button was controversial. I felt there were times when an immediate response was needed. So you push the button and speak. No writing. No speech-to-text. You say in your own words what you want. (Lots of problems here about nuance in text and non-verbal communication) Somehow that gets to the other person. Never worked that out.
At this point you have enough features that I'd argue you could support 80%+ of the activity people use the net for. Would you still need to research stuff? Sure. Play games? Sure. Converse in realtime about things you're interested in? Sure. But those are different physical units. You need to make both a mental decision and a physical effort to do those things. It should be apparent visually to yourself and others that these are the types of activities you are now engaged in. There can be no confusion, either internally or to an outside observer.
Tech is supposed to be about helping others. Instead it's become about making more tech. This was a fun exercise. Made me realize how far we are, and continue to go, from what we really need.
It may be an unappreciated market, but a lot of other folks have tried and failed. I got zero interest in this, aside from people who also had a problem with tech addiction. I believe the problem for most is: where's the payout? To be done correctly, this may work best as a philanthropic effort, not a startup. This isn't an under-appreciated market. This is the place with the greatest current gap between human suffering and people willing to help end it.
( This is when I started writing about the problem: https://news.ycombinator.com/item?id=6349447 I even made a crappy demo video as part of my first "real" F# project: https://vimeo.com/14460868 )