A lot of ram is under lifetime warranty, check the manifacturer site (usually a serial lookup is enough).

Yeah, Beaver City!

Personally I’m more in the “oh yes please take my job as a programmer” mindset, yet none of the currently available tools seem to be anywhere near capable of it.

When they do, I guess I’ll just translate the b.s. customers spit into something that’s even humanly readable, just merely distilling the intent often isn’t enough. Also A.I. needs to ‘learn’ to say no, because even though the customer asks for something, doesn’t mean they actually want to have the result of their question.

You could try clearing the cache in Heroic? It’s under settings -> Advanced

journalctl is usually the way yes, you can also check dmesg if it’s part of your distro. Errors are usually highlighted in red, that’s something to scan for. You can use journalctl -ef to keep the log updating on your screen, if the issue is intermittent, it might take a bit before it shows up. If the issue is logged on boot, it should be in journalctl -b

Does htop/free show the correct amount of memory?

Also try applying load on the system, maybe it’s a thermal or power delivery issue.

Do you have a overclock enabled in your bios? Try disabling that and XMP if enabled.

That’s a troubling state! My first guess would be your system memory being not great seated, which Linux handles more gratefully.

Do you see any errors in the Linux log?

Out of general curiosity, what hotel is it? asking for a friend ofc

Simplest solution would be to setup the nfs/smb as storage for backups and making a backup schedule. Datacenter -> Storage -> Add -> SMB/CIFS

Datacenter > Backup > Add

On Error Resume Next

Visual Basic is a beautiful language

“A bit loud” is understating it, those drives rip and tear (we use exos X18 drives). I pity the person trying to sleep next to those.

They are good though, while we had one (of 5) fail within the first week but that was quickly resolved.

All Linux iso’s, right?

A smart powerplug and/or a fingerbot would solve that problem I guess? But at that point it’s probably cheaper to buy a network connected picture frame.

You could use something like the Toshiba flash air?

Or the tld is .mobi

Systemd timer to poll upower when running on battery power, when battery is at 20%, use either system beep or set system volume and play a sound?

When a new game is released I usually check if it’s steam deck compatible, if it isn’t for no specific reason (like, a 2d platformer, I’m not going to expect a high fidelity 3d game to work) I’m way less inclined to buy it. The market is there and really should be picked up.

I’ve used it, it’s pretty rough and unfinished, the current main branch doesn’t build without help and you’ll need ollama or openai keys.

The results however are impressive, even with a small model like phi3 mini through ollama. They got some good prompts behind it and the results name the sources + have some good followup questions.

tbh, a lot of big players (Microsoft, Facebook, Google) host a lot of AI stuff on huggingface and quite likely have to pay for that.

Also they had a few successful funding rounds, last one led by Salesforce.

Also Amazon is invested in them, probably offering a lot to them for free or discounted.

that uses mDNS, which in some cases requires your router to be online to be able to resolve it to a ip. If part of your internet disruption was your router going down, it would explain the issue

How is your coordinator linked to your home assistant?

For example if you use Zigbee2MQTT and you have either Zigbee2MQTT or Homeassistant pointing to the internal ip of mosquito (192.168.1.11 for example) and your router goes down (with dhcp), it’s possible it cannot communicate anymore.

This isn’t the case if it’s all running on the same box using localhost as address, running it in a docker network or when you run ZHA however.