Unless you have a directional antenna, which these don't, you can not possibly scavenge power and simultaneously hope to increase the range by rebroadcasting.
So rebroadcasting is pointless, these simply communicate with the base station.
But what do they do? I mean it's nice to communicate, but you have to have something to say.
There are others doing work on this for sensors you could embed into say a wall to give you the pressure reading of the concrete (a simple piezo say) but those usually modulate an RF wave (similar to NFC) instead of rebroadcasting.
Sounds like 24 GHz. But all your critiques are still very valid.
But if you can blanket the world with these kinds of radios and mesh them, think 10s or 100s of trillions of radios, then the range issues are less of a concern but you have a chicken and egg problem which needs many many billions of dollars of investment to kickstart. Great reason for IPv6, though :)
1 - It works at about 60GHz. You are not getting a lot of distance or even going through a wall at that frequency and (likely) very low power levels
2 - The devices still need a central "base station". Presumably they rebroadcast to neighbouring nodes; but at lower power levels since they reuse some of the received power for themselves. I doubt you'd get more than 2 - 3 hops out before you don't have much power to begin with.
3 - He cites including these in fridges, coffee makers, light bulbs, etc. None of these are power constrained applications. You can put in $0.45 802.15.4 chip in them TODAY and it comes with a whole microcontroller inside to do what you want to do.
The tech here is no doubt "cool" but I don't see any applicability for IoT. If anything I think it would have some medical uses in pills you can swallow etc. But IoT use is kind of a dud.
Most IoTs have pretty good power source and you can run the w/ existing chipsets.