If I were a teenage geek, I'd do it for fun - especially if the neighbor was a friend. Likely, it would escalate to where we'd be turning each other's lights/home theatres/appliances/etc. on and off randomly. Just imagine... my neighbor loves classical music - so I change their streaming audio to an 'all Iron Maiden all of the time' channel. Oh the possibilities :-).
I think it is time to move to the middle of nowhere and get off the grid. I'll get rid of 'smart' meters with solar panels, generators, cisterns, and a well. I'll get radio/TV from the airwaves. If I want to watch a movie, I'll either drive into town and go to the theatre, or pay cash for a blu-ray. If I really need to get on the net, I'll find an internet cafe somewhere and post anonymously. Goodbye cell phones that follow you wherever you go. My car will be a '73 240Z.
I really don't like the idea of everyone from the utilities to the NSA to AT&T/Amazon/Google/etc. knowing my location and all my habits, likes, dislikes, etc. Worse, with the IOT the neighbors might hack into my home grid and turn on the furnace on a hot day just for fun.
The irony is that I helped design the technology that enables all this! Is it too late to repent?
I had a discussion with a friend of mine back in the 90's regarding the Internet appliances designs that we were building for our customers. This was the era of the "Internet-enabled toaster" and other bad ideas. My friend is a big data kind of guy, and he was telling me about XML and other standards that enabled the exchange of data between loosely coupled networked systems. That's when I realized that the real power of the Internet was going to be what is now called M2M communications. The problem at the time was that the infrastructure was not yet ready to support that expansion.
Fast-forward to today, and we are very close to being able to realize that expansion. There is real technology in place and useful things that can be done with it. Industrial applications do in fact do this type of communication, but it is done largely on an ad-hoc and local basis. This is similar to the Novell Netware networks that we used to build, which only reached within a company. Ultimately, IoT will juice the expansion the same way that TCP/IP did with the last one. Yes, there are issues. We must work out an appropriate security model. We will end up defining classifications for sensors as being openly public or restricted in terms of access. Nobody paid much attention to that in the Internet appliances because they never got serious enough to be subjected to that scrutiny. IoT is real enough and close enough that it is being examined on that level. The process will be messy and frustrating, but it will happen.