Home Assistant 2026.3: A clean sweep
Home Assistant 2026.3: A clean sweep
I wish I had some water meters that I could monitor to take advantage of the Energy dashboard, but sadly I don’t have a submeter I can access.
Home Assistant just keeps methodically getting better!
This update seems pretty tame from the release notes, but I was worried about the python update.
I have a lot of HACS integrations, and several HACS custom repository integrations that are updated very slowly when things get updated around them.
I’m happy to report that all of my integrations work just fine after the update.
Oh, and before I forget: have you seen our brand new merch store?
🤮
Why so negative? Bands also makes merch to increase their income, should we hate on them as well?
Oh, look at that: the Linux Foundation also sells merch linuxfoundation.store - maybe you should stop using Linux?
Microsoft also has meerch, so does Android.
I can’t find any merch store for Apple, so I guess your an Apple fanboy.
Why so negative?
It’s clear that the Nabu Casa Inc. people, who also happen to be the Home Assistant project leaders, are focussed on making money over making well engineered software.
For example, Home Assistant’s settings page includes an entry for Nabu Casa Inc.'s cloud services product as the first entry as the first entry in the list which cannot be switched off easily.
Home Assistant is engineered in such a way as to make it difficult to install on operating systems that aren’t under control of Nabu Casa Inc., like Home Assistant OS or Home Assistant Container. If Home Assistant were engineered well, it would be possible to take individual Home Assistant packages and compile and install them on any distribution, as has been customary in the free software community for decades. As far as I know, there’s no reason Home Assistant must be an operating system rather than simply individual packages. Using containers as a means of software distribution is bad engineering. See feddit.uk/post/17543373 .
Bands also makes merch to increase their income, should we hate on them as well?
If a band makes selling merch their purpose, over and above making music, then I would likewise scorn them.
It’s clear that the Nabu Casa Inc. people, who also happen to be the Home Assistant project leaders, are focussed on making money over making well engineered software.
It’s clear that Paulus created a free home automation product and developed it for 5 years. For free.
In 2018 they started a fund raising system which also helped provide secure remote access for those that don’t know how to do it themselves.
I’d say they are focussed on making well engineered software over making money.
I’d say they are focussed on making well engineered software over making money.
I think you must have a different idea of what “well engineered software” means because to me, nothing you’ve said implies a focus on making well engineered software.
Writing software without remuneration doesn’t imply a focus on well engineered software. A person can write software without remuneration with a focus on financial reward in future. Which is exactly what appears to have happened.
Helping secure remote access for those who don’t know how to do it themselves doesn’t imply a focus on well engineered software. Educating people isn’t the same thing as writing software, let alone writing good sofware.
This comment thread is about your opinion on the developers wanting to gain renumeration for their efforts.
Considering their influence on standardising open protocols across the home automation industry rather than proprietary lock-in protocols, then this is a much wider commitment than just software development… attending conferences is not cheap.
This comment thread is about your opinion on the developers wanting to gain renumeration for their efforts.
No it’s about the Home Assistant developers valuing profit over well engineered software.
I actually missed that on my first pass.
Thanks for the heads up!
I also love the breakdown of costs. It’s one of those little things that shows their mentality.
HA is fantastic once you’re past the learning curve but is still aggressively unintuitive sometimes, and I find little irritants all the time. Why can’t you make the Settings pages and sub-pages top level menu items? Why are entities and devices buried so deep in Settings? Why can’t you edit Zones from the Map view? Why can’t you easily rename entities in bulk on a per-device basis? Why can’t you automatically replace entity references in Automations with an updated entity name? Why isn’t there good documentation overall about how the system works instead of just technical documentation with narrow focuses? Why is the discord full of Linux elitist types who expect you to know the system when you’re trying to learn it?
It’s the little things everywhere that make me long for a Valve Software level of polish. There has been progress like the push to not manually configure things with YAML, but it’s so slow.
Dim the lights that are already on, and ignore the ones that are off?
I’m just pointing out here that you and I have different expectations; how could the software know what you intended?
99% of the time I want to adjust the current lighting, I don’t want to first turn on all lights and then adjust all of those lights to a uniform standard before individually toggling them all individually. Powering on all unpowered lights when adjusting brightness should be the edge case, IMO (also again not just my opinion, but the industry standard)
For the record all other smart home systems treat room groups the way I am describing (like a dimmer knob and power switches). But there isn’t even an option in HA for rooms to “only adjust devices currently in use”. The smart home companies seem to have researched how people naturally intuit such things.
If you’re adjusting the same lights repeatedly, you could set them up as a group.
It’s more work, but you could also write a script that detects the current status of each light then sets the brightness if it’s on. I use something like that for our smart porch lights that are on a smart switch - if the switch is off, turn it on, wait a bit for the lights to get on the network, then set them to the right color or whatever. (The switch normally stays on, but it gets turned off occasionally and it doesn’t automatically turn on after a power outage.)
I haven’t used other automation systems - I avoided them because I didn’t want to get locked into one, until I found HA. I also have never thought to “dim a room” - actually I’ve never used entire room controls at all, they never made a lot of sense to me, but then I generally only have one or two lights in a room to control.
The conversation topic was unintuitive aspects of HA, I’m aware hacky workarounds exist, but I find this (pretty central) behavior quite clunky.
I also find it crazy that you’ve never wanted to dim or brighten more than one light at a time lol but then again, diversity is the spice of FOSS!
I know how it works, I’m just saying it’s unintuitive. It’s not how any other smart home system works.
I use adaptive brightness too, actually. But nearly every time I’m manually adjusting a room’s existing brightness, I don’t want every single unpowered devices to turn on, too.
Flash with ESPHome? Not that I’m aware of (but would be nice to know)
I’m happy with the Shelly bulbs, they work without internet shenanigans and can do MQTT, etc…
I have used Pi Zeros and ESPresense for prsence detection… seem to work ok, but if you’re wanting to do room-level detection, (ie who’s where), then that seems to be really difficult to tune the sensitivity