Some of them have been implemented at small scale. I believe the "floor that turns foot traffic into power" has been implemented a couple of times. Certainly the "lift a rock to power a light" has been implemented several times. I believe someone even stuffed a soccer ball or three with things that generate power when the ball is kicked.
In fact it's a perennial design winner in various design contests that are judged by designers and not engineers or scientists. It looks awesome when you design some cool little gadget that "harvests" enough energy to power a small town from a soccer ball or something. The fact that it doesn't work is not a disqualifier in those contests. It't totally what I'd enter at this point if I were going to enter such a contest, it's an obviously winning play.
But the floor that converts foot traffic to energy doesn't generate much energy, and is very unpleasant to walk on. The soccer ball that extracts energy from the kick means that it has extracted energy from the kick, so it doesn't go flying away like a soccer ball should but is more likely playing with a ball of damp laundry. Guess what kids don't want to play with? The "lift a rock to power a light" does work, but it provides a lot less light for a lot less time than people think, and the rock ends up being pretty heavy. (But it does have at least some legitimate use cases, just not as many as you might hope.) The bike that extracts energy from the biker is so insanely unmarketable that what's winning in the market right now is the exact opposite, the electric bikes that supplement the biker's power.
(Also I will add as honorable mention the "solar powered X" that is spec'ed with one or two orders of magnitude too few solar panels. Putting panels on the top of your electric car may even be theoretically worthwhile in the long run, but only as a small supplement to the car. The idea that you'll be toodling down the highway at 75mph on 10ish square feet of solar panels is risible. Multiply that by some large two-digit numbers.)