I can get notification to ntfy, but I’m not sure if the app is certain to blow my phone up until I notice it, which is my goal. Frankly, if I could trigger the Presidential Alert, I would do that.

This is my summer solar system. I like to winter in the Antares.

I’m trying to figure out how to add it to Mailcow Dockerized and hook the existing containers. If I sort it out, I’ll probably PR it to Mailcow. I think it would a nice addition to start to build out a network that isn’t susceptible to the same spam attacks as regular email (yet).

I remember as a kid I set up one of the first private Echomail nodes as part of my RBBS bulletin board. UUCP was a big part of that, as I was the hop for other nodes coming onboard in my area. I added another half-dozen modems eventually just to handle the email traffic, then had to offload it to a university because I didn’t want to have to charge for the traffic and it was getting too big to handle. But it was pretty interesting at the time.

INtersting link to the LLM DDOSing against FOSS project sites/repos that’s been going on.

I love this line:

and robots.txt is the only thing they don’t read off the site

What a shitshow.

I’m using vscode/Roocode with Gosucoder shortprompt, with Requesty providing models. Generally I’ll use R1 to outline a project and Claude to implement. The shortprompt seems to reduce the context quite a bit and hence the cost. I’ve heard about Cursor but haven’t tried it yet.

When you’re using local models, which ones are you using? The ones I mention don’t seem to give me much I can use, but I’m also probably asking more of them because I see what Claude can do. It might also be a problem with how Roocode uses them, though when I just jump into a chat and ask it to spit out code, I don’t get much better.

That koboldcpp is pretty interesting. Looks like I can load a draft model for spec decode as well as a pile of other things.

What local models have you been using for coding? I’ve been disappointed with things like deepseek-coder and the qwen-coder, it’s not even a patch on Claude, but that damn cost for anthropic has been killing me.

Fair enough, but it’s damn handy and simple to use. And I don’t know how to do speculative decoding with ollama, which massively speeds up the models for me.

Man, I was excited about Technitium, but I’ve had a hell of a time trying to get it to work. I’m not sure if it’s intended to be on a DMZ in order to get TLS working or something, but I’ve not been able to get it to acknowledge a single DNS request, even when I think I’ve shut down DNSSec entirely.

Meshtastic node?

LMStudio is pretty much the standard. I think it’s opensource except for the UI. Even if you don’t end up using it long-term, it’s great for getting used to a lot of the models.

Otherwise there’s OpenWebUI that I would imagine would work as a docker compose, as I think there’s ARM images for OWU and ollama

If you click the HA header, there’s 5 different scripts there, each install HAOS in a different way. As a container in an LXC, as a “baremetal” LXC, as a VM, as a Pimox VM, or as a Podman container.

So there’s 5 methods in there of installing HA, not to mention having to take apart the scripts to figure out what they do in order to help you troubleshoot this.

Not sure what you expect of asking a question like this with no useful info about what you actually have in front of you.

I’ve quite come to look forward to this newsletter.

Just use the official Docker AIO and it is very, very little trouble. It’s by far the easiest way to use Nextcloud and the related services like Collabora and Talk.

If you use Live Migrate, realize that it doesn’t work on an LXC, only VMs. Your containers will be restarted with the LXC on the new node.

The advantages you gain with running a hypervisor on something like ZFS is immeasurable, for snapshotting, replication, snapshot backups and high availability. You don’t have to quiese machines to back them up and you can do instant COW snapshots before upgrades.

KVM doesn’t really have overhead, that’s the kernel part. Maybe a bit of RAM, but with LXCs it’s negligible.

1500 action minutes/mo limit.

deleted by creator

I like that. I tried to get Actions in Forgejo working and that was a dead-end. So I’ve been using act manually.

Appreciate the writeup.