If I understand correctly, developing the LLM's uses huge resources
and Nvidia GPUs. Once the models are developed, they can be downloaded
and run locally, with $10K worth of local hardware.

Also, the best free LLMs are approaching the best cutting edge stuff
from Anthropic and OpenAI.

It seems to me that businesses will want to have their LLMs local.
Scientists, coders and tech-savvy consumers will want local LLMs, too.

Ordinary folks will just type their questions into google, or maybe
a free version of Chat or Claude.

Graduate students at MIT etc. will amuse themselves developing specialized
LLMs, essentially forever.

If I understand this correctly, we may have a few boom years for hardware makers,
as a lot of $10K to $20K systems get sold to businesses, consumers, scientists,
and engineers.

It looks to me like OpenAI and Anthropic are screwed, unless they can follow
Microsofts model, and figure out how to make money of a product that is
available as a free download elsewhere.

Cheers,
Stopped Clock

But