cross-posted from: https://reddthat.com/post/76421889
Linux doesn’t deserve this
Imagine taking the lightweight, privacy-focused appeal of Linux and completely ruining it by hardwiring AI models straight into the OS. Nobody is switching from Windows for more forced AI features.
Arguably even worse. The kernel is Linux (99% human made), the desktop is probably based on something existing (probably 50% human made), but you can’t actually do anything. AI is integrated into everything. Think Windows but worse, somehow. Probably with less telemetry, at least! If the LLM is on-device, which it probably isn’t. It’s probably connected straight to ChatGPT APIs
This is funny to me because I highly doubt the LLM runs locally, which means the OS
becausebecomes unusable if there’s no internet.Edit: spelling
Exact opposite. KORA with Cortex is local first by default where “local” can be an on-machine LLM and / or ones that located on the same local network. So it’s Hybrid with local first and the added party trick that you can tap an LLM on another machine that is also running Cortex.
If it has local LLMs and other models it is gonna have gigantic minimum requirements.
Damn, the antithesis to Hanna Montana OS.
You get the worst, of both worlds
At the current state, I would say its way more than 99% human made. Probably like 99.99% human made, and the 0.01% generated by Ai is checked by the Linux experts, so no slop comes through.