Some of us remember TV remote controls that operated ultrasonically. They were terrible - somebody rattles their keys or the dog walks by with its metal collar clinking - lots of ultrasonic noise created by those events and Chunk-Chunk-Chunk you would be three channels away from your nightly news.

Seems solvable in this case. Ultra-sound can wake up the device, but then it would take some other signal to have it do anything other than go back to sleep, like WiFi/Bluetooth/LAN/etc.. So you wouldn't get incorrect functionality triggering from accidental ultra-sound triggers, you'd just save less battery.

My little brother used to do electrical installations. He once set up a device that deactivated the power in a building as soon as no more consumers were on. Of course that thing is probably still using power continuously itself.

On that same note, shouldn't it be possible to install a mechanism that doesn't use any power at all? Something like, uh, I'm not gonna say a light switch, but a weightless switch, to mix a few metaphors here.

It consumes 8 nanowatts, which is well below the self-discharge rate of almost all batteries.

only nanowatts? With an antenna, could this low level of power be drawn from radio/TV signals? It is enough to power a crystal set radio receiver.

Nanowatts are -60 dBm or what you would expect to receive on your cellphone when you have decent but not great signal. Many radios operate down to the -110 dBm/MHz PSD or lower ranges for high performance stuff. So you could use hits those power levels with energy harvesting on radio waves but only if you were in the right location, on the right band, and it would possibly cause problems for the systems that are trying to use those signals for their intended purposes.

Yes, a passive wake up receiver is possible. See https://ip.sandia.gov/technology.do/techID=175

An RFID tag would do the job.

This isn't novel. As an idea, signal amplification being used for switching has been around forever. Plus this thing still draws power. A completely passive receiver is possible.

Yes, I was expecting a receiver that draws power from the received signal.

It's the security and privacy implications that come to mind for me.

Care to elaborate how this applies? The article is about a switch. The alternative to a switch is to leave the device on full-time, or implement some self-resuming feature using a RTC. Both of those require constant power. I don't think any of this has anything to do with 'security/privacy implications', so some examples would be appreciated!