ITT: “Haha, yah AI makes shitty insecure code!”

<mad scrabbling in background to review all the code committed in the last year>

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.

Until we got agriculture.

Well, see, your mistake is brushing your teeth and living past 30. If your back molars were properly rotten enough to gracefully pop out when the wisdoms grew in, and then you died before that one rotted and you couldn’t chew anymore, you wouldn’t have any problems.

Literally.

Check the returns policy, but if you could buy a large external drive at something like BestBuy, do your copy then return it, that might be a lot safer than what you’re talking about. Just a thought.

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.

Windows Docker is so bad, I don’t even know why it’s a thing.

Some good planning might make the migration less painful. I would recommend a ZFS or other COW storage solution under the docker host so you can do snapshot backups and not have to worry about quiesing databases, etc.

You might also want to check out https://yams.media/, it’s pretty much an install script and configuration walkthrough that’s very complete and detailed. Includes most relevant Arrs and gluetun builtin. Containerized. Choice of Emby, Plex or Jellyfin.

Chuck Testa

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.

What’s your throughput to another Windows box from your source machine? Samba isn’t known as a fast transfer protocol. You could also try NFS services on the windows box and export your shares to try that.

You might also enable jumbo frames if it’s a lot of larger files.

ARC will give way to other memory uses. Were you actually seeing memory shortage or did you just see the usage and think that you couldn’t overprovision?

https://superuser.com/questions/1137416/how-can-i-determine-the-current-size-of-the-arc-in-zfs-and-how-does-the-arc-rel

Meshtastic node?

Claude is the standard that all others are judged by. But it’s not cheap.

Gemini is pretty good, and Qwen-coder isn’t bad. I’d suggest you watch a few vids on GosuCoder’s YT channel to see what works for you, he reviews a pile of them and it’s quite up to date.

And if you use VScode, I highly recommend the Roocode extension. Gosucoder also goes into revising the roocode prompt to reduce costs for Claude. Another extension is Cline.

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

AIO is performant and much easier to maintain. If there was a method to try to run Nextcloud in the last decade, I probably tried it, and nothing has compared to the AIO.

Yikes.

Thanks for posting that.

Oh, ok. Every picture I saw made it look like a bowden extruder sitting on the end of the X gantry, with the tube going over to the hotend.