Then helmfile might be worth checking out

again.

On a serious note: space is hard and I wish them the best of luck for their next attempt.

You dont need to manually handle the WG config files. This isn’t really an issue when it’s just you and your two devices, but once you start supporting more people, like non-technical family members, this gets really annoying really quickly.

Tailscale (and headscale) just require you to log in, which even those family members can manage and then does the rest for you. They also support SSO in which case you wouldn’t even have to create new accounts.

The steam page has a demo that features the full game but has saving disabled.

Highly recommend taking that for a quick ring-dive.

Thanks for that handy search result link.

Really makes me want to play Rings of Saturn again :)

What an incredible image.

I almost like it more than the artist rendition, even though it is way easier to understand/visualize.

Is this a real “photograph” (including non-visible or even radar imagery) or computer generated from a simulation of some sort?

Regardless of the sponsorship in this video, SuperfastMatt’s videos are awesome. Really interesting projects delivered with great humor.

There are some experimental models made specifically for use with Home Assistant, for example home-llm.

Even though they are tiny 1-3B I’ve found them to work much better than even 14B general purpose models. Obviously they suck for general purpose questions just by their size alone.

That being said they’re still LLMs. I like to keep the “prefer handling commands locally” option turned on and only use the LLM as a fallback.

Thats what I’ve been using as well. On some of my cards it has some weird layout bugs (only on some viewing devices) which annoy me.

What card are you using for your room overviews?

I’ve done the same. I put on a single layer of clear coat and it has been perfectly water tight.

Mostly its really fucking expensive. Usual applications are central control of heating and window blinds in large office buildings.

Part of what drives the price is that ideally it’s all hard wired.

Opening the app for the first time on my Fairphone 5 (listed as unsupported) actually crashed the OS, but after that it seems to be working ok.

Closing out of the in-app gallery causes the app to crash. But that can easily be worked around by using some other gallery app.

I’ll be testing it for a bit to see how it fares against other HDR methods…

Why not set up backups for the Proxmox VM and be done with it?

Also makes it easy to add offsite backups via the Proxmox Backup Server in the future.

This person had the same issue and they’ve just logged out and in again

Additional information regarding Home Assistant:

The sun component (which should be enabled by default) already computes the sun position for you.

Elevation and azimuth are available as standalone sensors sensor.sun_solar_azimuth (might be disabled by default) or as attributes on the sun.sun entity.

I don’t have any experience with it but this might do something along those lines(?):

https://esphome.io/components/binary_sensor/ble_presence.html

Seems like you can just add it to one or more of your existing esphome devices.

If you have such a system up and running already you could try to modify it before ripping it out and starting from scratch.

Borrowing an idea from the machine learning approach you could additionally take the difference in average outside temperature yesterday and the average forecasted outside temperature today. Then multiply that by a weight (the machine learning approach would find this value for you but a single weight can also be found by hand) and subtract it from the target temperature before the division step discussed previously. Effectively saying “you don’t need to heat as much today since it will be a little warmer”.

I fear that’s about all you can do with this approach without massively overcomplicating things.

This is effectively what a thermostat does.

The problem is that the controller won’t know how well insulated each room is, how cold it is outside (including wind speed), which doors and windows are open and when, what people or devices are doing in each room.

The way thermostats solve this is by creating a closed loop where they react to how the room reacts to their actions.

Depending on how your heaters work you’ll likely need some dynamic component to react to these unforeseen changes unless you can live with the temperature being very unstable.

To get a rough idea of how long the heaters will have to run you can look at each room in for the last n days and see if the heater’s runtime was long enough to (on average) hold your target temperature. Dividing the average temperature with the target temperature will give you an idea whether they were on for too long or too short. (If the heaters have thermostats you’ll likely need to subtract a small amount from that value so that it will settle at the minimum required heating time)

If that value is close to 1.0 you know that on those days the heating time was just about perfect.

Once that is the case you can take the previous days heating time and divide it up over the cheapest hours. The smaller of a value n you choose the more reactive the system will be but it will also get a little more unstable. Depending on your house and climate this system described here might simply be unsuitable for you because it takes too long to react to changes.

There are many other ways to approach this very interesting problem. You could for example try to create a more accurate model incorporating weather and other data with machine learning. That way it could even do rudimentary forecasting.