9 ms·

I dunno, I find it pretty hard to get on board the home automation train. I feel like, working in tech, we know better than anyone else how incomplete or unreli

I dunno, I find it pretty hard to get on board the home automation train. I feel like, working in tech, we know better than anyone else how incomplete or unreliable these systems can be, so why introduce it to your home? Why replace your plain old metal kettle with a fancy smart-kettle that you can turn on with your Alexa, but would easily crap out on you if your Internet goes down, or some circuit breaks, or Amazon's API starts being flaky, or you lose electricity in your home? Shouldn't we be a little more worried about introducing this kind of extra complexity into our lives? (^ these aren't meant to be sassy rhetorical questions, I'm genuinely curious if everyone else just views this a different way than I do)

spiderfarmer 7y ago
Not GP, but for me it’s a small hobby, I want to do more with computers than just designing, developing and running websites. I use homekit/homebridge to do some basic stuff, switching on/off all the lights, computers, displays and audio gear in my office all at once, pushing just one button. This saves me a minute each time I enter and leave the office and it’s just very convenient This setup has worked flawlessly for a full year now and I don’t expect it to break any time soon.
wayoutthere 7y ago
There’s another part that people forget: analog tech is nigh bulletproof. Your $50 smart light switch may last a decade or more, but a $0.50 analog light switch will last a century. The extreme difference in cost often leaves you with something that is worse in the long run. Agree about the complexity as well. Analog circuits don’t have an API that changes every few years — a modern 3-way light switch circuit would be easily understood by an electrician from 1920 (and vice-versa). Good luck doing that with Lutron gear from a decade ago. I probably dropped $5k on home automation stuff. While I could afford it and it was partially a research project, the main lesson I learned was that home automation is generally more trouble than its worth at any price. Sure, it’s great that I can tell Alexa to turn my lights off, but I also have to make sure that the 3 services required to make that work remain talking to each other. I would cumulatively save more time by just walking downstairs to turn the lights off.
- ghaff 7y ago
  I use Alexa for one light that never got connected to a light switch. I used wireless X-10 for it previously :-) I agree about reliability. It's one reason I won't get a Nest thermostat. I live in snow country and I'm not going to do anything that potentially introduces a new failure mode into my heating system. (I actually have a couple of the Nest smoke detectors. The internet connection isn't required for them to function as dumb detectors and they work well in my experience.)
  - jld 7y ago
    An internet connection isn’t required for the Nest thermostat to operate the furnace either. I can appreciate not wanting to allow Google the opportunity to brick your thermostat (and hence your furnance) through some kind of software issue but if you’re willing to extend that faith to their smoke detectors, why not your furnace?
    - ghaff 7y ago
      I guess my feeling with the smoke/CO detectors is that I also have other "dumb" models throughout the house. Furthermore, a thermostat needs to work basically all the time at certain times of the year, whereas the detector only needs to work when there's something to detect. I guess Google could silently brick them and I'd never know about it but that seems to be getting into paranoia territory. (And it's not like other detectors can't silently fail too.) It's probably paranoia about the thermostat too TBH. But other than looking cooler than my old programmable Honeywell, a Nest thermostat also wouldn't really do anything for me anyway.
      - wayoutthere 7y ago
        Yeah, the Nest was ahead of its time in 2010; but by 2012 the big players like Honeywell and Carrier had caught up. At this point they beat Nest on both price and features. A good friend of mine owns an HVAC company and would recommend a dozen thermostats over a Nest on functionality alone.
  - 7y ago
melling 7y ago
I don’t see where he used a kettle. A contrived example? I don’t own one myself because I use a microwave. How about the stuff he discussed in the article? All of it seems valid. Personally, I wouldn’t mind a little home automation that works with a voice ui. Over time, it’ll become cheaper, more reliable, and transparent. Also, I prefer to optimize for 99% of my time instead of the rare event. I can pull out a pot when the power goes out. Like people who would rather use gas cars instead of electric because of range and there are 180,000 gas stations in the US, you might want to check back in 10 years.
iamaelephant 7y ago
> Shouldn't we be a little more worried about introducing this kind of extra complexity into our lives? Yes, we should. I'm currently building and I wouldn't touch this stuff with a 10 foot pole. The only thing I'm considering is Philips Hue bulbs, but only for the configurable colour temperature and brightness, and if I can find an alternative that doesn't need a web connection I will use that.
- Havoc 7y ago
  >Philips Hue bulbs, but only for the configurable colour The xiaomi equivalent is pretty good too (yeelight) at half the price. That's assuming your OK with the brand >alternative that doesn't need a web connection I will use that. All these things need an initial Web connect and usually app to configure. After that you can firewall it off and use a local server like home assistant. Well for the xiaomi. Not sure Philip's
  - iamaelephant 7y ago
    Great tips, thanks mate I'll look into that.
  - drdaeman 7y ago
    Philips Hue doesn't require Internet connectivity. Bulbs and switches use ZigBee ZLL (AFAIK, they don't have WiFi capabilities at all), and IIRC it can even work without the bridge (just turning on and off) as switches can talk directly to bulbs. Bridge device surely can work without Internet. It's essentially a small webserver that exposes an API for the smartphone app (which talks to the bridge directly as long as they're on the same network) and then talks to the devices over ZigBee. IIRC, there is even FLOSS RPi-based alternative implementation somewhere out there. Don't know if Internet is required for the initial setup.
- trashcan 7y ago
  Lutron Caseta is really good for smart switches. Also, if you just want motion activated lights, there are switches with IR motion sensors built in.
maxerickson 7y ago
I'm considering setting up proximity/motion lights for a relative that has mobility issues. Centralizing the control has some nice advantages over just using motion switches, and they would likely be in addition to the existing lighting, not replacing it. The idea there would be that they have the light when they need it, without having to switch them on and off.
luma 7y ago
The trick with a well-implemented smart home system is to consider failure modes in the design phase. I made my own "smart-kettle"[1] which does work with Alexa but when the Internet is down it reverts to a regular kettle with regular buttons. Similarly, smart lights are often problematic for a range of reasons. Most need to use an app to control them, or some incantation for Alexa to do the same, neither of which are discoverable (or maybe even accessible) to someone in your house who isn't you. If you have them connected to a regular light switch, they stop being "smart" when someone uses the light switch on the wall to turn them off. All of these problems are easily addressed by using smart switches instead of smart lights. Everybody knows how to use a light switch, and any failure of your "smart" systems means the light switch still works like every other light switch out there (meaning, get off your butt, walk over to the switch and it turns lights on and off). This concept holds true in my house for garage door openers, door locks, fireplace controls, and a whole raft of other stuff. The existing systems function like they should to a user who knows how a garage door opener (or whatever) is traditionally supposed to work. In addition, they get some extra functionality[2] if everything is up and working as designed. If things aren't working... well you have to press the garage door opener button like it's 2010. [1] https://hackaday.io/project/20217-iotea-kettle https://hackaday.io/project/20217-iotea-kettle [2] https://github.com/aderusha/MQTTCarPresence https://github.com/aderusha/MQTTCarPresence
- zabzonk 7y ago
  > I made my own "smart-kettle" Not obvious what fills it with water.
- Slartie 7y ago
  > All of these problems are easily addressed by using smart switches instead of smart lights. This is so true. I really wonder about all the people installing these WiFi enabled light bulbs into lamps that end up being connected to some regular dumb switch next to the door in the wall...what's your rationale for putting the "smarts" in the bulb and not the switch, the device whose sole purpose has always been to stop and start power distribution to a dumb bulb that just glows when it gets powered? The failure mode of augmented switches, if the smart control stuff for some reason doesn't work or is not accessible, is super obvious and simple: they just revert to being a switch on the wall. And it's not just about that, it's also often just more convenient to enable the light the old-fashioned way by touching a rectangle when entering a room. I'm very satisfied with my light setup, allowing control of the living room lights via programmed conditions as well as by a Logitech Harmony remote. But most other members of my family tend to ignore these and just like to use the switch. Sometimes I do as well, like when entering or exiting the room, but I wouldn't want to miss the additional remote control options. With the smarts in the switch (and the home control server behind it wiring it all together), we're all satisfied.
jpmattia 7y ago
> I feel like, working in tech, we know better than anyone else how incomplete or unreliable these systems can be, so why introduce it to your home? I think there is an S-curve argument in here somewhere. There was a time where putting in an electric stove was just asking for trouble because of the reliability, but there were early adopters who were excited about the idea. Fortunately, those guys shook out the bugs for the generations that followed. Ultimately though, I'm probably in the same part of the S-curve as you. I spend much of my day doing (hacking | improving | debugging) (code | business | people | self) so adding another layer sounds like a PITA, but if you're an early adopter, my hat is off to you and I send my thanks for shaking out the bugs for the rest of us!
- bryanrasmussen 7y ago
  Our stove broke towards the end of 2018 after nearly 20 years. I went to buy a new one, really was difficult to buy anything that wasn't an electric one. So I spent money on getting one of the better ones because I didn't want any problems. It's been broken and needed a repairman to show up twice since purchase.
ip26 7y ago
You can just do little things. I bought some WeMo smart outlets that plug in to regular sockets- basically a WiFi lamp timer. They have been a pleasant upgrade to old fashioned dial lamp timers, easier to change the schedule, trigger with sunrise/sunset, etc. You can combine them with all sorts of appliances. For example, I'm always irked how much idle power my garage door opener draws. Well, I could throw a smart switch inline to shut it off at night, and in the event of late night door usage I can override the schedule with my phone. If they were to fail or EoL, I'd be out some money but all you have to do is unplug them.
- weberc2 7y ago
  I do the same, but I do worry about the security aspect. I don’t like the idea of having a bunch of devices on my network. I put the IoT things on a guest network, but I still feel iffy about it. I’m sure there are better solutions, but all of them seem to require flashing your router with DDWRT and doing fancy networking things that I don’t understand well, and frankly I don’t want to put the effort in.
- Scoundreller 7y ago
  > how much idle power my garage door opener draws How much does it draw? Standard Chamberlain or something exotic?
  - ip26 7y ago
    I do believe it's a Whisper drive Chamberlain, I measured it somewhere around 15W around the clock. $10/year or so. You might think I'm just unlucky, but it turns out most of them are like that.
Jnr 7y ago
Don't use Amazon/Google/other service then. Make it work locally in your intranet. (And optionally reach your intranet with Wireguard VPN if it is desired.) A lot of smart home devices support ZigBee standard and with a device like ConBee (ZigBee USB adapter) you can get rid of the reliance on external services. Run your own gateway that can talk with a lot more different devices and run it in the middle of nowhere, if necessary. You can also run your own voice command recognition software if you also want to have some voice control for all of the devices.
michrassena 7y ago
I had similar thoughts last night. As a person who's spent almost 25 years in IT, I don't see any compelling reason to make things which "just work" into a personal project. Assuming a person were able to get everything tied together and working perfectly, you've still built a contraption at the mercy of capricious third-parties. You've committed to an unknown amount of future effort into replicating existing services or adding gee-wiz features which look futuristic but don't add value. There is still much to go wrong in an ordinary light switch, but I have plenty in my home over 50 years old which work perfectly. Where I live, I can flip the switch and the light comes on and I don't even have to think about it, often forgetting I had done so. It is this mindlessness and lack of attention which defines convenience, which purports to be what home automation is about. I can't imagine what kind of revolution must take place in the computer industry to create smart home automation products with equivalent cost, longevity, convenience, and which instill equivalent confidence in their operation.
pfranz 7y ago
I'm skeptical of adopting for adoption sake, but I've dipped my toes in where other solutions wouldn't work. I've kept in mind the failure mode (or hacker exposure). I also like monitoring/reducing consumption as a bit of a hobby. I previously had a dark stairway to the house and got home after dark. A timer would have sufficed (I haven't seen them for wall switches), but I got a IoT switch that turned the light on at sunset. I had a detached garage that let out to a busy street and would often forget to close it when driving away. I didn't get around to installing something, but wanted something to close it after 15m of being open. My floor heaters consume a lot of power and other people in the house forget to turn them off. When they were installed, they included WiFi and an app. I wrote a bridge to HomeKit and set it to turn off at 11am and 11pm. When traveling I reduce the heating/cooling. It's nice (especially when getting home close to midnight) to turn it back on when landing at the airport so you don't come home to an uncomfortable house. It's not quite IoT, but my kid is too young to tell time so we have an alarm that changes color when its ok to wake up. I had a dumb one, similar to an alarm clock. It was annoying to set. A more expensive, bluetooth, one sets time from my phone. I was a bit worried about the bluetooth one when traveling (time zone changes, not having wifi), but it worked great. I think the nicer one worked better because of build quality and not because of IoT stuff. I don't have any "scenes." Just a few, very specific, items with rules. I added outdoor Christmas lights this year because last year my kid kept hiding the remote and I didn't plug things into an accessible place outside. A lot of people seem to like being able to turn off house lights without getting up from their bed. I'd never blindly recommend any of this, but if it might solve a problem they have, I'd walk them through my experience.
BrandoElFollito 7y ago
I was thinking about that recently, having automated many lights at home. This was done in an experimental way, with an amateur approach, and ended up failing to work when I was away and leaving my wife with 30% of the lights only. It was not a big deal for me since I was away :) But surprisingly not that a big deal for her either, after explaining me how she would fillet me when I am back home. But then she read more than usual (the tv was out of service as well) and it was finally ok. My lights usually works but the added value of the automation is really there. Sure we can live without but it is a nice touch. I would say that YMMV depending on how you balance availability with some added value you could live without.
clairity 7y ago
no need to be sassy, you can add smart home products that are useful, not prone to singular failure, inexpensive, and privacy-preserving. i've installed a dozen or so zigbee-based ikea tradfri led bulbs and switched outlets in my apartment (along with a half-price zigbee-to-wifi gateway from the bargain bin) and they work great. if any part of the network goes out, i can just use them as regular bulbs by actuating the switches already in the wall or built into the lamp. my favorite addition is using a motion sensor for the walk-in closet bulb, since the built-in light switch is awkwardly placed in a back corner. i open the door and the light goes on, close it and it goes off a few seconds later, a simple pleasure. the tradfri gateway also now connects via homekit to my existing apple tv, which is used as the apple home server. that way, i can turn lights on and off from the control center on my phone from anywhere via icloud. the apple tv and gateway are on their on vlan and have restricted access to the internet and the rest of my network. sure, you can go crazy trying to add automation to everything, but so far, i've had no extra stress or complexity, just simple additional convenience.
tlrobinson 7y ago
I make it a rule that all "smart" devices should function first and foremost as "dumb" devices, and the "smarts" should only augment their function. A light should never require a smartphone and app to turn on and off, but it may require it to change the color, for example. I also try to buy devices with local control, vs. cloud-based control.
joe_the_user 7y ago
Yeah, it seems like the physical actions around an ordinary house are just too simple and discreet for automation to be desirable. Analogue, immediate action have physicality that is satisfying to a our being; take out a pot, fill it with a cup of a water, add half a cup of rice, light a burner and begin the process. Sure: "give me rice!" might seem simpler but "let us now debug the rice program" will become as difficult and infinitely less satisfying. See: Ready-at-hand and all the Heideggerian terminology concerning this stuff.
- icebraining 7y ago
  Curious example; while typical home automation is rare around here, kitchen robots (particularly Thermomix) sell like hotcakes. People are happy to pay a full month's wage to debug rice programs.

spiderfarmer 7y ago

Not GP, but for me it’s a small hobby, I want to do more with computers than just designing, developing and running websites. I use homekit/homebridge to do some basic stuff, switching on/off all the lights, computers, displays and audio gear in my office all at once, pushing just one button. This saves me a minute each time I enter and leave the office and it’s just very convenient This setup has worked flawlessly for a full year now and I don’t expect it to break any time soon.

wayoutthere 7y ago

There’s another part that people forget: analog tech is nigh bulletproof. Your $50 smart light switch may last a decade or more, but a $0.50 analog light switch will last a century. The extreme difference in cost often leaves you with something that is worse in the long run. Agree about the complexity as well. Analog circuits don’t have an API that changes every few years — a modern 3-way light switch circuit would be easily understood by an electrician from 1920 (and vice-versa). Good luck doing that with Lutron gear from a decade ago. I probably dropped $5k on home automation stuff. While I could afford it and it was partially a research project, the main lesson I learned was that home automation is generally more trouble than its worth at any price. Sure, it’s great that I can tell Alexa to turn my lights off, but I also have to make sure that the 3 services required to make that work remain talking to each other. I would cumulatively save more time by just walking downstairs to turn the lights off.

ghaff 7y ago
I use Alexa for one light that never got connected to a light switch. I used wireless X-10 for it previously :-) I agree about reliability. It's one reason I won't get a Nest thermostat. I live in snow country and I'm not going to do anything that potentially introduces a new failure mode into my heating system. (I actually have a couple of the Nest smoke detectors. The internet connection isn't required for them to function as dumb detectors and they work well in my experience.)
- jld 7y ago
  An internet connection isn’t required for the Nest thermostat to operate the furnace either. I can appreciate not wanting to allow Google the opportunity to brick your thermostat (and hence your furnance) through some kind of software issue but if you’re willing to extend that faith to their smoke detectors, why not your furnace?
  - ghaff 7y ago
    I guess my feeling with the smoke/CO detectors is that I also have other "dumb" models throughout the house. Furthermore, a thermostat needs to work basically all the time at certain times of the year, whereas the detector only needs to work when there's something to detect. I guess Google could silently brick them and I'd never know about it but that seems to be getting into paranoia territory. (And it's not like other detectors can't silently fail too.) It's probably paranoia about the thermostat too TBH. But other than looking cooler than my old programmable Honeywell, a Nest thermostat also wouldn't really do anything for me anyway.
    - wayoutthere 7y ago
      Yeah, the Nest was ahead of its time in 2010; but by 2012 the big players like Honeywell and Carrier had caught up. At this point they beat Nest on both price and features. A good friend of mine owns an HVAC company and would recommend a dozen thermostats over a Nest on functionality alone.
- 7y ago

melling 7y ago

I don’t see where he used a kettle. A contrived example? I don’t own one myself because I use a microwave. How about the stuff he discussed in the article? All of it seems valid. Personally, I wouldn’t mind a little home automation that works with a voice ui. Over time, it’ll become cheaper, more reliable, and transparent. Also, I prefer to optimize for 99% of my time instead of the rare event. I can pull out a pot when the power goes out. Like people who would rather use gas cars instead of electric because of range and there are 180,000 gas stations in the US, you might want to check back in 10 years.

iamaelephant 7y ago

> Shouldn't we be a little more worried about introducing this kind of extra complexity into our lives? Yes, we should. I'm currently building and I wouldn't touch this stuff with a 10 foot pole. The only thing I'm considering is Philips Hue bulbs, but only for the configurable colour temperature and brightness, and if I can find an alternative that doesn't need a web connection I will use that.

Havoc 7y ago
>Philips Hue bulbs, but only for the configurable colour The xiaomi equivalent is pretty good too (yeelight) at half the price. That's assuming your OK with the brand >alternative that doesn't need a web connection I will use that. All these things need an initial Web connect and usually app to configure. After that you can firewall it off and use a local server like home assistant. Well for the xiaomi. Not sure Philip's
- iamaelephant 7y ago
  Great tips, thanks mate I'll look into that.
- drdaeman 7y ago
  Philips Hue doesn't require Internet connectivity. Bulbs and switches use ZigBee ZLL (AFAIK, they don't have WiFi capabilities at all), and IIRC it can even work without the bridge (just turning on and off) as switches can talk directly to bulbs. Bridge device surely can work without Internet. It's essentially a small webserver that exposes an API for the smartphone app (which talks to the bridge directly as long as they're on the same network) and then talks to the devices over ZigBee. IIRC, there is even FLOSS RPi-based alternative implementation somewhere out there. Don't know if Internet is required for the initial setup.
trashcan 7y ago
Lutron Caseta is really good for smart switches. Also, if you just want motion activated lights, there are switches with IR motion sensors built in.

maxerickson 7y ago

I'm considering setting up proximity/motion lights for a relative that has mobility issues. Centralizing the control has some nice advantages over just using motion switches, and they would likely be in addition to the existing lighting, not replacing it. The idea there would be that they have the light when they need it, without having to switch them on and off.

luma 7y ago

The trick with a well-implemented smart home system is to consider failure modes in the design phase. I made my own "smart-kettle"[1] which does work with Alexa but when the Internet is down it reverts to a regular kettle with regular buttons. Similarly, smart lights are often problematic for a range of reasons. Most need to use an app to control them, or some incantation for Alexa to do the same, neither of which are discoverable (or maybe even accessible) to someone in your house who isn't you. If you have them connected to a regular light switch, they stop being "smart" when someone uses the light switch on the wall to turn them off. All of these problems are easily addressed by using smart switches instead of smart lights. Everybody knows how to use a light switch, and any failure of your "smart" systems means the light switch still works like every other light switch out there (meaning, get off your butt, walk over to the switch and it turns lights on and off). This concept holds true in my house for garage door openers, door locks, fireplace controls, and a whole raft of other stuff. The existing systems function like they should to a user who knows how a garage door opener (or whatever) is traditionally supposed to work. In addition, they get some extra functionality[2] if everything is up and working as designed. If things aren't working... well you have to press the garage door opener button like it's 2010. [1] https://hackaday.io/project/20217-iotea-kettle https://hackaday.io/project/20217-iotea-kettle [2] https://github.com/aderusha/MQTTCarPresence https://github.com/aderusha/MQTTCarPresence

zabzonk 7y ago
> I made my own "smart-kettle" Not obvious what fills it with water.
Slartie 7y ago
> All of these problems are easily addressed by using smart switches instead of smart lights. This is so true. I really wonder about all the people installing these WiFi enabled light bulbs into lamps that end up being connected to some regular dumb switch next to the door in the wall...what's your rationale for putting the "smarts" in the bulb and not the switch, the device whose sole purpose has always been to stop and start power distribution to a dumb bulb that just glows when it gets powered? The failure mode of augmented switches, if the smart control stuff for some reason doesn't work or is not accessible, is super obvious and simple: they just revert to being a switch on the wall. And it's not just about that, it's also often just more convenient to enable the light the old-fashioned way by touching a rectangle when entering a room. I'm very satisfied with my light setup, allowing control of the living room lights via programmed conditions as well as by a Logitech Harmony remote. But most other members of my family tend to ignore these and just like to use the switch. Sometimes I do as well, like when entering or exiting the room, but I wouldn't want to miss the additional remote control options. With the smarts in the switch (and the home control server behind it wiring it all together), we're all satisfied.

jpmattia 7y ago

> I feel like, working in tech, we know better than anyone else how incomplete or unreliable these systems can be, so why introduce it to your home? I think there is an S-curve argument in here somewhere. There was a time where putting in an electric stove was just asking for trouble because of the reliability, but there were early adopters who were excited about the idea. Fortunately, those guys shook out the bugs for the generations that followed. Ultimately though, I'm probably in the same part of the S-curve as you. I spend much of my day doing (hacking | improving | debugging) (code | business | people | self) so adding another layer sounds like a PITA, but if you're an early adopter, my hat is off to you and I send my thanks for shaking out the bugs for the rest of us!

bryanrasmussen 7y ago
Our stove broke towards the end of 2018 after nearly 20 years. I went to buy a new one, really was difficult to buy anything that wasn't an electric one. So I spent money on getting one of the better ones because I didn't want any problems. It's been broken and needed a repairman to show up twice since purchase.

ip26 7y ago

You can just do little things. I bought some WeMo smart outlets that plug in to regular sockets- basically a WiFi lamp timer. They have been a pleasant upgrade to old fashioned dial lamp timers, easier to change the schedule, trigger with sunrise/sunset, etc. You can combine them with all sorts of appliances. For example, I'm always irked how much idle power my garage door opener draws. Well, I could throw a smart switch inline to shut it off at night, and in the event of late night door usage I can override the schedule with my phone. If they were to fail or EoL, I'd be out some money but all you have to do is unplug them.

weberc2 7y ago
I do the same, but I do worry about the security aspect. I don’t like the idea of having a bunch of devices on my network. I put the IoT things on a guest network, but I still feel iffy about it. I’m sure there are better solutions, but all of them seem to require flashing your router with DDWRT and doing fancy networking things that I don’t understand well, and frankly I don’t want to put the effort in.
Scoundreller 7y ago
> how much idle power my garage door opener draws How much does it draw? Standard Chamberlain or something exotic?
- ip26 7y ago
  I do believe it's a Whisper drive Chamberlain, I measured it somewhere around 15W around the clock. $10/year or so. You might think I'm just unlucky, but it turns out most of them are like that.

Jnr 7y ago

Don't use Amazon/Google/other service then. Make it work locally in your intranet. (And optionally reach your intranet with Wireguard VPN if it is desired.) A lot of smart home devices support ZigBee standard and with a device like ConBee (ZigBee USB adapter) you can get rid of the reliance on external services. Run your own gateway that can talk with a lot more different devices and run it in the middle of nowhere, if necessary. You can also run your own voice command recognition software if you also want to have some voice control for all of the devices.

michrassena 7y ago

I had similar thoughts last night. As a person who's spent almost 25 years in IT, I don't see any compelling reason to make things which "just work" into a personal project. Assuming a person were able to get everything tied together and working perfectly, you've still built a contraption at the mercy of capricious third-parties. You've committed to an unknown amount of future effort into replicating existing services or adding gee-wiz features which look futuristic but don't add value. There is still much to go wrong in an ordinary light switch, but I have plenty in my home over 50 years old which work perfectly. Where I live, I can flip the switch and the light comes on and I don't even have to think about it, often forgetting I had done so. It is this mindlessness and lack of attention which defines convenience, which purports to be what home automation is about. I can't imagine what kind of revolution must take place in the computer industry to create smart home automation products with equivalent cost, longevity, convenience, and which instill equivalent confidence in their operation.

pfranz 7y ago

I'm skeptical of adopting for adoption sake, but I've dipped my toes in where other solutions wouldn't work. I've kept in mind the failure mode (or hacker exposure). I also like monitoring/reducing consumption as a bit of a hobby. I previously had a dark stairway to the house and got home after dark. A timer would have sufficed (I haven't seen them for wall switches), but I got a IoT switch that turned the light on at sunset. I had a detached garage that let out to a busy street and would often forget to close it when driving away. I didn't get around to installing something, but wanted something to close it after 15m of being open. My floor heaters consume a lot of power and other people in the house forget to turn them off. When they were installed, they included WiFi and an app. I wrote a bridge to HomeKit and set it to turn off at 11am and 11pm. When traveling I reduce the heating/cooling. It's nice (especially when getting home close to midnight) to turn it back on when landing at the airport so you don't come home to an uncomfortable house. It's not quite IoT, but my kid is too young to tell time so we have an alarm that changes color when its ok to wake up. I had a dumb one, similar to an alarm clock. It was annoying to set. A more expensive, bluetooth, one sets time from my phone. I was a bit worried about the bluetooth one when traveling (time zone changes, not having wifi), but it worked great. I think the nicer one worked better because of build quality and not because of IoT stuff. I don't have any "scenes." Just a few, very specific, items with rules. I added outdoor Christmas lights this year because last year my kid kept hiding the remote and I didn't plug things into an accessible place outside. A lot of people seem to like being able to turn off house lights without getting up from their bed. I'd never blindly recommend any of this, but if it might solve a problem they have, I'd walk them through my experience.

BrandoElFollito 7y ago

I was thinking about that recently, having automated many lights at home. This was done in an experimental way, with an amateur approach, and ended up failing to work when I was away and leaving my wife with 30% of the lights only. It was not a big deal for me since I was away :) But surprisingly not that a big deal for her either, after explaining me how she would fillet me when I am back home. But then she read more than usual (the tv was out of service as well) and it was finally ok. My lights usually works but the added value of the automation is really there. Sure we can live without but it is a nice touch. I would say that YMMV depending on how you balance availability with some added value you could live without.

clairity 7y ago

no need to be sassy, you can add smart home products that are useful, not prone to singular failure, inexpensive, and privacy-preserving. i've installed a dozen or so zigbee-based ikea tradfri led bulbs and switched outlets in my apartment (along with a half-price zigbee-to-wifi gateway from the bargain bin) and they work great. if any part of the network goes out, i can just use them as regular bulbs by actuating the switches already in the wall or built into the lamp. my favorite addition is using a motion sensor for the walk-in closet bulb, since the built-in light switch is awkwardly placed in a back corner. i open the door and the light goes on, close it and it goes off a few seconds later, a simple pleasure. the tradfri gateway also now connects via homekit to my existing apple tv, which is used as the apple home server. that way, i can turn lights on and off from the control center on my phone from anywhere via icloud. the apple tv and gateway are on their on vlan and have restricted access to the internet and the rest of my network. sure, you can go crazy trying to add automation to everything, but so far, i've had no extra stress or complexity, just simple additional convenience.

tlrobinson 7y ago

I make it a rule that all "smart" devices should function first and foremost as "dumb" devices, and the "smarts" should only augment their function. A light should never require a smartphone and app to turn on and off, but it may require it to change the color, for example. I also try to buy devices with local control, vs. cloud-based control.

joe_the_user 7y ago

Yeah, it seems like the physical actions around an ordinary house are just too simple and discreet for automation to be desirable. Analogue, immediate action have physicality that is satisfying to a our being; take out a pot, fill it with a cup of a water, add half a cup of rice, light a burner and begin the process. Sure: "give me rice!" might seem simpler but "let us now debug the rice program" will become as difficult and infinitely less satisfying. See: Ready-at-hand and all the Heideggerian terminology concerning this stuff.

icebraining 7y ago
Curious example; while typical home automation is rare around here, kitchen robots (particularly Thermomix) sell like hotcakes. People are happy to pay a full month's wage to debug rice programs.