[ExI] How new AI economics might change our systems

[John Clark]

*Mr. Gemini wrote:*

*> "While the commenter’s diagnosis of structural change is compelling,
> declaring that we can simply "throw all previous economics out of the
> window" introduces serious blind spots.The Physical Bottlenecks (Bits vs.
> Atoms).  AI may have near-zero marginal cost in software, but it operates
> in the physical world" *


*That would be a valid criticism except for the fact that as LLMs keep
getting smarter robots are simultaneously becoming more dexterous. And on
that topic I was particularly impressed by a new AI that was introduced
just two days ago called "Jev" that I think would be perfect for robotics.
Jev is different from all other AI's, it doesn't use transformers and it's
not even a large language model, it doesn't output text, instead it outputs
decisions based on probabilities and degrees of certainty. Jev is much
smaller and cheaper to operate than other AIs and is hundreds of times
faster.  And it has a zero hallucination rate, so although it could still
be wrong because it works on probabilities, it will never make a decision
based on a fact that it made up as LLMs sometimes do. *

*A new kind of AI model from a ChatGPT inventor is thrilling developers*
<https://techcrunch.com/2026/09/18/a-new-kind-of-ai-model-from-a-chatgpt-inventor-is-thrilling-developers/>

*Mr. Gemini** also ignores Drexler style Nanotechnology which AI will
certainly speed up the development of; and once we have that the
distinction between bits and atoms starts to become very blurry, at least
when it comes to manufacturing. *

*John K Clark*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20260920/6cb44549/attachment-0001.htm>