I run on dual Xeon R410s with 128Gb of RAM (2013?). Got them for free, on Kijiji. Runs Proxmox on both and a pile of VMs. Dual GB nics, 6 SAS bays, HBA in IT mode for ZFS. Has iLo for OOB management, or whatever the Dell equivalent is.
I mean, it’s not fast, but each server has 24 cores and I can chunk PDF files fairly quickly for RAG on 10 cores and have plenty for mail server, Nextcloud, K8S running some side hustle apps, etc, etc. Kind of a noisy prick when it winds up though.
Thanks for the feedback. That was precisely my worry about outlaying that money and not being happy with the result.
I’ve really been mulling one of those over with 128GB. I’m on Claude Max and Cerebras $50 so I’m using a good amount of $200/mo for coding and Openclaw. Is it worth it for light coding, or are you only doing SD with it?
This was the channel I was going to suggest. A lot of what he shows is pretty pricey, but some would make sense if you weren’t too concerned about speed.
Framework desktop?
I built a docker wyoming version of pocket-tts a while ago because I like the voices and timbre of pocket-tts better than Piper. I’ve subbed it in for Piper.
Also been using the OVH Linux-voice-assistant which I dockerized for voice hardware on my desktop.
I would use their LXC install, it’s much more flexible. It does not need to be local but it does simplify things like email. I had to put a bit of effort into getting it to be able to connect to IMAP mailboxes to process, but it wasn’t any more than just asking it to get the necessary libraries etc. But things like that are why using it as an LXC is a better choice. It might be able to do that as a docker, but there’s potential problems with network connectivity and docker in docker issues.
You can also firewall that LXC off without having to mess up your own workstation, as well as snapshot it and back it up.
And the first thing I would do is have it keep token budgets when you build tasks, and report it’s token use to you every hour or two. It takes some time to learn how to structure reminders and task processing to not create loops that eat up scads of tokens. Don’t ask me how I know.
But holy hell, can it be useful.
zfs.rent
I had a pig named Porkchop.
She had great taste.
I set up Pulse recently and the ease of setup and great UI/UX is impressive. Really liking it.
Of course, there’s some AI bullshit if you want to opt in, but it’s not enabled by default.
I tried it a couple years ago and it wasn’t very successful. But maybe that’s changed.
The other server sends again until it times out? Never been an issue, that’s just how email works. Most SMTP servers will attempt for a few days if they know the MX is valid. Besides, I’ve never been out for longer than an hour or two.
Seems really low. I run it on a docker LXC with Nextcloud and bunch of other stuff, on 8 cores of an ancient dual Xeon Dell. I never seem to have to deal with latency on it.
Mailcow-dockerized. I’ve used it for nearly a decade, it’s been flawless. Very easy to set up with the admin webpage and has a webmail client, or use Roundcube with it.
Make sure you have your DKIM, SPF and Dmarc records in order and tested against MXtoolbox before you start.
Frigate is painful to set up. It won’t just go out and query the onvif capabilities so you have to try and figure out its RTSP url manually and ptz support is primitive. Its low resource and stable once you manage to get it to work.
Blue Iris is much easier and more capable, but uses Windows, and its a resource hog, and its paid. But if you get past that, BI is really good.
That’s a real bitch if you’re running watchtower and have the latest tag set. If you aren’t watching this drama, it’s an easy pwn for whoever took it over.
Docker Hub needs to get their shit together.
I do OK with faster-whisper for transcribing, and I built a wyoming container for pocket-tts that does pretty good local TTS but I am running it as a docker on a ryzen machine (no gpu) so YMMV. Pocket-TTS seems better than Piper IMO, it’s certainly faster for local TTS if all you’re using is CPU.
I’ve been looking for a MCU that has enough oomph to do some noise cancellation onboard and I ordered up a couple of these in the hopes that the onboard NPU would be useful for that. It also has a speaker output and onboard mic. Price was right for 8GB of EMMC and 256MB of ram.
Oh, people will keep using it no matter how much you warn them.
Proxmox-helper-scripts is a perfect example. They’ll agree with you until that site comes up, and then its “it’ll never, ever get hacked and subverted, nope, can’t happen, impossible”.
Wankers.