Everybody is talking about LLMs, aka Large Language Models, sometimes via the cutesy word Llama, which means the same thing except smoothed over and rounded off by the linguistic optimization of getting used in actual speech.
I don’t know this for sure, but I suspect that the falsely-modest adjective “Large” came about because a more emphatic term, like “Massive,” would lead to the unfortunate acronym MLM.
It’s already the case that the “language” part of LLM is getting outdated. New models, and most of the exciting things we’ll see in 2024, are “multimodal,” fluent not only in language but also in pictures, speech, songs, video, and even movement through 3D space.
The most likely outcome is that we’ll be stuck with the term LLM long after it meaningfully describes the tools we work with.
But I’m hopeful that we’ll be able to pick something more accurate…I’m just not sure what that might be. “Large Multimodal Model” is practically a tongue-twister.
One other option worth considering. If “language” is no longer the operative word here, what about “learning”? (which is the essence of neural networks, on which these things are all based) Then we don’t even have to pick something other than LLM / Llama.
(This post 100% human written)
reposting from my LinkedIn comment:
some say the destiny of AI (or at least, one of the main applications) will be a small furry “animal” that you carry on your on your shoulder, where it will work as your copilot/wingman. So I suggest using either of these terms instead Llama/LLM.
In addition: the Sci-Fi author, Matthew Hughes, in his Henghis Hapthorn series, provides the main character with a device called an “integrator” that acts like a hyper-intelligent familiar. So I suggest using the term PI to stand for “personal integrator”. Simply written as the symbol ‘π’.
Just my 2¢