Instead of that I'm choosing to vote with my wallet and mostly stay away until this is resolved. Skyrocketing inflation is not doing anything to change my mind either.
Instead of that I'm choosing to vote with my wallet and mostly stay away until this is resolved. Skyrocketing inflation is not doing anything to change my mind either.
I've settled on a ZigStar Stick for my always-on Raspberry Pi and stick to compatible devices[0] only. I even just stick to the dashboard of Zigbee2MQTT[1], as Home Assistant[2] feels too bloated for my use cases.
Self-hosting and open source will certainly win in the case of smart homes.
[0]: https://www.zigbee2mqtt.io/supported-devices/
[1]: https://flemmingss.com/wp-content/uploads/2021/09/zigbee2mqt...
Or at least only trade away long term reliability when there are useful features being gained.
Eternally relevant rant about smart product solutioneering: https://m.youtube.com/watch?v=HcXu4_K1tMQ&t=45s
Stop pestering me because you think I haven't given you enough money yet. Go away.
But it is available as a plug-in appliance, and they're getting better and better at the Just Works aspects all the time.
And in a world where everything is relative, it's a hell of a lot easier to get started with Home Assistant than it is to get started with something like KNX or Crestron.
I have all kinds of widgets around me at home that are the products of successful companies of various sizes, and that are so niche that 95% of people would probably never seek to own such widgets themselves. Most probably aren't even aware that widgets like this exist in their universe. But the remaining 5%? They know about the widget, and they want one (maybe more than one).
There's ~149 million households in the US alone. 5% household penetration would require ~7.5 million widgets.
It sure is easy to imagine ways that a person and/or a company can sell just 7.5 million widgets and ultimately be financially unsuccessful, but it's also easy to imagine ways in which that would work out just fine.
But that's just meandering digression. :)
The Open Home Foundation had revenue of CHF 8,836,394 and expenses of of CHF 7,352,798 in FY 2025[1].
Is that small potatoes in a world with trillion-dollar companies? Yeah, it's certainly not ginormous. Is that successful? I personally think that it is.
I'd much rather have an open version of Apple Home or whatnot, even if I paid for it. I really don't want to spend a bunch of time dicking around with home automation - I want it to simply do what I want it to do.
That Matter light bulb still needs something to control it, but the light bulb is not tied to one control ecosystem -- at all. It works with any Matter-supporting system.
That bulb can even work multiple different Matter systems concurrently. Which sounds dumb, but that helps smooth our transition periods wherein a person switches from one system to another. The old system can continue to operate just fine while the new system is implemented, and then once the new hotness is behaving Good Enough the old one can just be switched off.
There's also complications around networking. Matter over WiFi works great if your networking gear can withstand potentially hundreds of devices connected. Matter over Thread seem to work under specific conditions, but I still haven't been able to pair the first device.
There's also potential issues around device certification.
All in all, Zigbee is a more reliable platform for the moment IMO, even though I have some issues with it as well.
I get all of the features I think I'd want out of something like a light bulb or a wall switch that just responds to commands and reports status. :)
None of my stuff uses Thread yet. I'm cheap, so I tend to buy home automation gear at the very low end of the market and Thread just isn't quite there yet.
Anyway, Zigbee.
I was using Zigbee when that still mostly bare dev and demo boards, >20 years ago. It was neat then (I was using it to consolidate RS-485 serial comms between a bunch of gear mounted on different radio towers), and it's still neat today.
But in home automation, Zigbee has been problematic for me, with devices going offline and coming back according to some unseen agenda. I've gathered that "right way" to fix that is apparently to use repeater-capable devices here or there, but meh: I'd rather build a proper wifi network (with hardwired access points) than buy weird stuff I don't need, put it in places where it isn't otherwise useful, and let Zigbee form its own unsteered mesh.
It just makes more sense to me to build a single wireless network, than to build multiple wireless networks. So I built a proper-enough wifi network with some Mikrotik hardware and that's what I use.
But I still have one Zigbee device (a temperature sensor) and it's been working fine for years. I don't track the rate at which it eats coin cells as well as I could, but I probably get a little over a year out of a random Chinese cell and they don't cost much.
In terms of wireless bandwidth: I find it only sensible at this point to consider 2.4ghz to be trash. It's not worth preserving. Internet speeds (and needs) keep increasing, and that means that 2.4ghz decreases in relative usefulness for work that requires high speed data. And all of my serious devices have 5ghz radios, anyway, and have had them for a long time.
Its biggest strength is also its biggest limitation: Penetration through walls tends to be better, and that means that the band gets stuffed up with noise from the neighbors more than 5ghz does.
It seems OK to make phone calls using a Bluetooth headset at short ranges, though. And it's plenty good-enough to tell a light bulb to turn on or off, run a music streamer, or play with esp32 and pi pico projects.
So even though many others seem to despise the concept: I'm quite deliberately centered on using 2.4ghz wifi for home automation stuff. It works well. I think I reached this position rationally. :)
Interestingly enough, I agree that cheap Matter over WiFi devices work well. I have a bunch of Moes lightbulbs and they worked out of the box, except one that seems to lose its connection after a few minutes of pairing - still pretty good out of about 30, each costing £3-4.
The bigger brands are the worst offenders - SwitchBot, Meross and Tapo are all problematic to an extent.
I'm also just building a solid WiFi network for these devices. If nothing else, I can troubleshoot WiFi a lot more easily than either Thread or Zigbee.
Unfortunately you see quite a few manufacturers implement Matter in a way where they require a proprietary hub, and the proprietary hub requires a constant internet connection to work...
In what way?
I don't see any of them that require these things.
If you do see them yourself, then please use the space provided by clicking the "Reply" button to name-and-shame. Please be specific; I need this information in my life.
If I'm looking at that right. it looks like Switchbot sells a device that talks Matter on one end, and emits infrared commands on the other end.
And that it needs continuous Internet access in order to perform this infrared function because it doesn't store those infrared codes locally -- like, apparently, not at all.
Every programmable remote for the past 30+ years has managed to do this with the most basic of hardware, but a company with the technological prowess of Switchbot just can't quite get that functionality into their workhorse ESP32.
That's hilarious. 10/10. I needed that today.
I'll try to remember to amend my future comments about Matter's real-world offline utility with "Unless they're the product of a resolute fuckup, like Switchbot."
That's good. I don't want a single company to control everything in my house. Having multiple vendors with different implementations is healthy for the industry.
The real problem is there's nothing "smart" about those "smart home devices" to begin with. They are just remotely controlled devices. Technically, someone or something still has to control them because these devices can't really make decisions on their own. And then, to make those decisions remotely (either in the cloud or in a home local controller), every home installation needs to be highly customized. There are several light switches in my house that I never need to touch -- their states are automatically shifted based on like ~15 conditions: time of day, light conditions, occupancy, human activities(iOS focus modes). Non technical people can't do that.
I would want those devices to make simple decisions on their own (I have a few "dumb" light switches with motion+light sensors and they are pretty good already), and for complex decisions, they expose some APIs for other devices/controllers to call. Matter seems to be the right direction, but it is still controlled by a handful powerful players.
Imagine that every vendor would have different implementation of TCP/IP stack.
But this seems not to be a well-advertised way of doing things, unfortunately.
Not only that, but within a few years they all stop functioning and now all you have is a collection of expensive paper weights.
The greed of the manufacturers made "smart home" a predictably dumb proposition.
And half of them completely disable local network access... Like why?!
Matter is all fine but try making a matter device as a hobbyist, you quickly realize how it was designed for large companies and people simply choosing simpler solutions that they can copy and paste practically.