I’m nearing completion here: I’ll write it up and post it when I get home and then it’s yours to pull apart use :)
PS: I personally use MCPProxy as a single endpoint for all MCPs; dunno if that helps you with orchestration at all?
I found this informative - YMMV.
From description:
The internet is incomprehensibly big and filled with slop. In this video we talk about how it got like that, how search and recommendation algorithms were designed to fix it, and why they are now breaking down.
I remember throwing out a Sony trinitron with my Dad when we got our first LCD. The thing must have weighed 80kg. It sat on the verge for God knows how long before it disappeared.
This would have been about 20 years ago. What a mistake.
Then again… I remember my folks throwing out or donating my old computers (C64, Amiga 1000, 500, 600). War crimes, I sez.
Good man/woman/inanimate carbon rod
You might like this
https://www.youtube.com/watch?v=o7IVcp3312Q
I’m just showing the kids episode 1 of Ren and Stimpy…wish I had a CRT for it
Yeah I can see both Flatpak and App Image
Seems like it’s there.
I’m coding a Kiwix mcp (I assume you’re using kiwix as your wikipedia source?) that might be of interest to you. I’ll post about it on !localLlama when done.
I don’t like any of the extant Kiwix MCP tools that I found because they leave formulation, ranking, search and interpretation up to the LLM.
I’m looking at a different way to get exact spans with re-ranking via TinyBERT right in front of the llm so that it doesn’t have to do 3 jobs and hallucinate itself into stupid, which is what small LLMs tend to do when given multi-hop tools.
I’m testing it with Ling Tiny 3, miniCPM5-2b, Granite 4-3b and Qwen 3.5-2b and 4b, with the goal of making sure that it actually generalizes. It’s been a massive pain in the ass and I’ve basically proven RAG from first principles…but when I’m done, you won’t need to vectorize the source into any kind of database nor “just trust me bro” the llm is searching Kiwix properly.
Well, if you’d rather a ready made solution or two
Like the poster above you, I’m rolling my own. More work…but a lot cheaper
PS: you know about this one, right?
https://shop.m5stack.com/products/m5stack-core2-esp32-iot-development-kit?variant=35960244109476
While I’m not the biggest fan of this (for technical reasons), this is quite simple to onboard
I mean, it’s pretty much a “one-click install your own local AI chatbot” at this point.
Because the context is likey to be “… to a third party” of a presumably human kind. Human to clanker (which then gets eavesdropped) ought not to qualify.
Ofc letter of law vs spirit of law, variation in jurisdictions, legal interpretation, YMMV etc etc
Exactly. And it was gobbling up 30 gigabytes / is switched on by default.
#JustMicrosoftThings
I literally just now disabled hiberfil.sys on Windows (it’s basically useless tbh) and reclaimed 30 fucking GB of space.
powercfg /h off
30GB.
WTAF.
$20 for a Compaq CQ45 or slightly newer on eBay…throw in a small SSD.
See my comments above as to what you can run with it. TL;DR a surprising amount
Can confirm that worked well into the Win8.1 era.
Heh. I have a core 2 duo (Compaq CQ45, circa 2010) with 4gb running Zorin Lite OS. For 99.99% for every day tasks (and then some) it runs excellently, with about 600-800mb in reserve before caching. We’re talking YouTube, LibreOffice, multiple Firefox tabs, GIMP…all it needed was a cheap ssd to replace the spinning rust.
For context: this thing struggles to run Half Life 2 at any sort of reasonable speeds (though that’s an iGPU issue to be fair).
Last glance this laptop costs $20AUD to purchase off eBay.
Win 10 LTSC, Tiny10 or Tiny11 no good for your brother’s use case? I haven’t kept up with the Tiny cooks but IIRC Tiny 10 used a comically small amount of ram.
This any good?
They went from $0 to 2x$0.
Outrageous, really.
PS: It’s a joke; user above is making a joke about plex vs jellyfin price increases.