Thoughts, learnings and regrets after three years on Home Assistant

Just a little blog of my own. Most of it will be familiar to people here.


Thanks for the share and I share your point of view about current state of HA .
Myself coming from closed automation systems market (AMX, Crestron,…), HA is so much better even with its little drawbacks but still can find a way around most of time thanks to open source ! (Interesting to note also that AMX, one of the leader in the past of home/conference room automation systems has released a new generation of controlers that is an expensive ersatz of HA :rofl:

I’m in the process of rectifying this same issue:

When I look back, there are a few things that I would do differently. I might have leaned more toward smart switches and integrated wall sockets than smart bulbs

Smart bulbs are fine for table lamps and indication but for room lighting when I have guests that don’t understand that all the lights are automatically controlled they need to be able to operate physical switches.


I really like the Hue bulbs – they put out much better light (much warmer, better colour) than the averge LED. I’d like to combine the approaches – have smart switches that just turn off the bulbs (but leave them powered).

There are many high CRI LED bulbs (> 90) no need for a smart one for that.

Thanks for sharing that. Good points, all.

I want to add one observation: You started as many of us have; looking for a good use case to try out a Raspberry Pi.

Fortunately, HA isn’t tied to the RPi or any other hardware. This is a huge positive.

The negative, IMHO, is a lot of us who use HA are born tinkerers. We love buying bigger, faster and “better” hardware. Of course HA can scale up to just about any environment we can imagine and build.

My concern is that some people on this forum suggest HA should be run on more expensive hardware. I’ve seen this pattern often throughout a long career in IT: Developers get the best hardware, then develop things which require that hardware to work effectively.

My plea is simple: Don’t abandon the users who are attracted to HA because it can run on a RPi. By doing so, you remove one of HA’s key “selling points.”

1 Like

I fully agree with this. I started with HA on a Rpi3b/1Gb ram with 7" touchscreen that I had lying around, which I was running in supervised and using as both the engine and the control/display unit. That worked, but it was really stretching the hardware. Then again, I might not have started using HA if I would have had to invest in a more expensive setup. All I wanted in the beginning was a thermostat that could run two different weekly schedules every other week. I only had to buy a relay. But the rabbit hole was opened…

I wrote a quick follow-up post on this, just FYI. I had cut out several things from the first one, just to get it published. These are just the bits I was going to mention as well.