In my experience (and one look at the Home Assistant forums shows I’m not alone), allowing AI to write Home Assistant configs is a questionable activity.
There’s so much old, wrong information out there that it regularly uses (or tries to use) deprecated or removed features, and often even hallucinates config fields if you ask it to do something that is unsupported (and, of course, the user doesn’t know beforehand what’s supported). And you better have backups, because once it’s inserted unparsable nonsense or messed up the white space in the YAML config files, Home Assistant just won’t work.
Admittedly, my experience here is six months old, and I was talking to Claude not using an agent - but still…
My heat pump had also thrown an error, but because it sits in a closet (the indoor part anyway), and we don't use its tiny screen for slow TV, nobody noticed until it started getting cold. To make things even worse, the heat pump resets its error messages every 24 hours and retries, removing the error message in the process, so you may check on the heat pump and it shows zero error, yet 30 minutes later it calls for technical service.
I ended up installing an eBus shield in mine. It is a €50 ESP32 device that reports data to MQTT, and HA can pick it up [1] from there.
Things went a little overboard from there, so there's smart electricity meter monitoring, water meter monitoring, automatic valve controllers on main valves, basically anything "house" can be monitored and controlled from HA.
On the bright side I have very few smart light installations. Lights are more or less a solved problem in my book. 100+ years of electricity has pretty much revealed the optimal placement for light switches, and while some lights truly benefit from smart bulbs or relays, the majority doesn't (in my home).
5-10 years ago I didn't care if smart technology had local control, today I won't purchase anything that doesn't allow me to control it locally in some fashion. I don't mind having to tinker a bit with ESPHome or MQTT, that's fine, and for the vast majority of people, a cloud integration is probably the right solution, but I want the ability to operate my devices completely off grid if need be.
Either AI wrote this, or the author thinks this is what humans sound like now.
This doesn’t make much intuitive sense to me because shouldn’t output based on statistics have a very strong through line?
This is where I really first got lost:
“You can’t run a greenhouse without gas unless you can see the moments it slips — a fault, a silent fallback — and close them one at a time. The monitoring doesn’t undermine the gas-free promise. It’s what makes the promise keepable.”
>"This is not bad luck. It’s a structure."
>"We just don’t control it. The vendor does."
Clearly written by an LLM, so many tells.
The empty phrases are one thing, but LLM articles like this seem almost perversely designed to take longer to read then they did to write.