LEDs don't transmit through walls unless they happen to be transparent, like, say, a WINDOW. And while your GHz-based wireless signal drops off just down the block, the signals leaking through your window can be picked up from miles away with a telescope. Didn't we learn this from the modem-blinkenlights exploit like 10 years ago? I suppose most of us have also learned how to use encryption in the interim..

I'm pretty sure this has already been done. I saw it in Total Recall like 15 years ago.

Squeezebox would be my choice too.. There's just something cool about hijacking all that unused copper for good use (plus it's dirt cheap).

Thanks for the lesson in basic electricity, I must have slept through that one. As noted below, the resistance of most phone lines is enough to prevent most amps from setting your children on fire. Obviously some caution should be applied when overloading your very expensive amp, but most amps will just distort or at worst blow a fuse if you drive them too hard. In my own experience, a $30/50W amp was able to drive two speakers directly plus 4 more distributed throughout the house without distortion, destruction, or death of any kind (impedance came out to ~3Ohm, IIRC). Yes, I had to turn the volume down on the amp to prevent distortion, but the output was plenty loud enough to fill the house.

The line-level solution would be great if our phone lines were shielded. I suspect this would just end up producing unbearable noise..

If you have stopped using your landline like most people I know, you can attach your amp to the phone lines and small speakers anywhere in your house that's near a phone jack. You need a reasonably sturdy amplifier if you want to drive more than 6 speakers or so, and the sound quality won't be dazzling, but it's real cheap and easy. You can also attach a cheap potentiometer to each speaker for volume control.

My #1 reason: if Linux had 15% desktop share, then hardware manufacturers could no longer ignore it and more software developers would consider supporting it.

Sensory substitution is old (but cool) news, but from TFA it looks like this guy is claiming some inherent ability of the skin to detect light, rather than delivering an image-driven stimulus to the skin. If this is the case, then he's got a lot of work to do. Like stop running simulations and start checking premises.