Hacker News Reader

Comment by gojomo | Hacker News Reader

The writeup describes Wattvision as being "whole-house only". For all I know, that's true.

However, if it's got sufficient resolution to give "15 second" updates, it seems you could train it to recognize certain draws: holding everything else constant, turn individual items on at exact cues given by the software.

(The software might even be able to form hypotheses about individual draws before they're exactly named. Then the user just goes in and recognizes certain steps on the constructed timeline: "oh, yeah, that's the hallway light I turn on when I get home from work, and that's obviously the electric dryer, and this random thing must be the refrigerator compressor coming on and off through the night.")

View on HN

You're right -- there's a lot of stuff Wattvision can do with the data we gather today. There's no reason we can't make smart conclusions about what's running (as gojomo suggests) or or eventually expand to per-device sensors. We're just starting simple.

Combine those two ideas and you can do "Your washing machines energy profile has changed for the worse, perhaps you should get it serviced?"

Wattvision sounds cool, but I'd be worried about their business model since PG&E are replacing their meters with digital ones that report back to PG&E. I seem to remember hearing that they are rolling out a similar service, too.

Any startup should consider itself blessed to compete with PG&E. (And if Wattvision's tech is superior, the big utilities are far more likely to become customers than competitors.)

There is also existing competition that does some REALLY cool things (like Tendril, http://www.tendrilinc.com/)

Why would that be a blessing? If PG&E comes and puts in a digital meter for which their little gadget doesn't work, they're back to page 0.

PG&E is a big, slow, regulated utility unlikely to rush in executing anything. And even if PG&E did rush to digitize, that would only mean more pressure on other utilities (even in other markets) to adopt a cheap ready-to-go solution, and more opportunities for software innovations that can take the data from any source (digicam or utilityco meter).

I'm sure that a lof of the people who this could help already know about it.

If you're in Queensland, Australia, the Climate Smart Home Service costs $50 and you get: - a wireless energy monitor - up to 15 compact fluro lightbulbs - a water/power efficient showerhead all with installation.

http://www.climatesmarthome.com/

Reply on news.ycombinator.com