Local LLM'S, Hardware, UI and Models

[TylerAbeoJordan]

I think we are all going to have to wait for a better hardware:

Groq AI's LPU: The breakthrough answer to ChatGPT's GPU woes?

(further in the future):

Watch this video!

History, explanation of, and possible future application of memristors. [Memristors for Analog AI Chips] So possibly after creating LLM (conventional way), the LLM could be encoded into a memristor. This super low power and high speed device could thus replace conventional high power LLM and make these intelligences available anywhere in any kind of device – possibly like an upgradable sim card, but for AI. Maybe could continue to learn as well.

and/or better LLM structure:

What’s up today? (Part 2)

@dirvine 1-bit LLM’s being developed by Microsoft – LLM’s radically reduced in size and computational (inference) complexity, without compromising performance. LLM’s on smartphones will be easily possible: [image] [What is 1-bit LLM?]

… and what I think would be the Coup de grâce for speed and efficiency - stacking highly specialized (and so vastly smaller) LLM’s – with a language interpretation layer and a logic layer as the top two layers – wherein the logic layer can access any number of other specialized LLM’s to get the skill or data it requires to serve the request.