- Posted on
Sun Feb 25, 2018 6:42 pm
-
dduff617
offline
-
- Posts: 661
- Joined: Jul 05, 2006
- Location: Massachusetts, USA
I personally have never found Nest "learning" to be worthwhile in any way -- not even a little bit. My situation may be unique, but I suspect that others may also have in common some of these reasons why (to me) it seems ill-conceived:
1. My thermostats are not in located in good positions to tell "home / away" so at best Nest would be working with weak information.
2. I live in a house with multiple people moving around on different schedules.
3. I have multiple heating zones in my house and I generally do not want all the zones to automatically switch to "Home" mode any time someone is home - that's way too simplistic. Instead I want "occupied" status to apply selectively to different zones at different times.
4. Its just not plausible to think an "AI" is going to guess right from a few times that I sent the temp at certain times what it should do on a recurring basis. Even if there were an intelligent person crammed inside, there's no way for him to make good guesses based on such noisy information. I could try to remember to turn the heat down several nights in a row at the same time and hope Nest would "learn" this, or I could just set up a schedule and tell Indigo to turn down my thermostat every night at the same time. The latter just generally seems better to me.
5. Nest can at best learn a simplistic "default" program, but AI isn't going to learn things like travel, school vacations, times when someone stays home sick, or other special situations anyway, so you're going to want to do some control yourself anyway - to set up "modes", controls, schedules, and triggers for these kinds of things. So what's the point of creating your controls to do precisely what you want some of the time and then just letting some AI do random stuff the rest of the time? That doesn't make work for me.
6. I'm not cloudaphobic, but I still don't feel great about a remote system outside my control making a bunch of inferences about when I am not likely to be home. Sadly I realize that's probably happening anyway, even though I've turned off the learning part.
7. I've gone to the trouble of having a local HA control system (Indigo) and I prefer to keep control there where I have all the sensors and other devices in my Indigo universe at my disposal. For example, if I were inclined to add some extra "smarts" to my HA system to better predict when I'm home (like geo-fencing, motion detection, door sensors, devices being on, etc.), I'd like to benefit from other "smart" uses of this rather than just having them buried inside the black-box of Nest's "AI".
On the other hand, I realize that others may be happy to go all in with the Nest Universe and "works with Nest" stuff. It will probably grow and get smarter over time. Nest makes cameras and security systems, so I could see their home/away detection getting much better than it is for me in my house currently, at least for those that continue to buy into their product line. If you lived in an apartment, didn't have a separate HA system, and really didn't like the programming interface of your $15 generic thermostat, then Nest may be a better fit.
All that said, if you turn off the "learning" part, Nest still makes a decent thermostat for my use with Indigo. It's controllable and communicates better/faster and more more reliably than my old Insteon/Venstar units did. Being WiFi based, it doesn't put any traffic at all onto your Insteon network, and since thermostats tend to be "chatty", that's probably a good thing for avoiding collisions and general "overcrowding" (if you have a big home and lots of devices). Their app is nice enough and interacts smoothly with the Indigo plugin - i.e., you can use other the app or Indigo to access the device and any updates are visible to both interfaces. They measure and report humidity. They qualify as "programmable WiFi thermostats" which for my local utility means there seems to always be a $100 rebate on them (I probably would not have bought them otherwise).