Category: AI, Computation, and Technology

I mentioned those technologies because they are all currently in some state of development, with, I think, next month being the first large scale neuromorphic computer coming online, I think in Australia? (pretty sure). Photonic chips, and three dimensional chips are all in progress. IMO we will see the boom in existing vector compute chips last for a bit and then a crash as those chips are replaced by the neuromorphic. But as in all things I have a very good record of WHAT will happen but timing such things is a fool’s game. 😉

Reply addressees: @RunicSigil

—“Q: What about mentats instead of chips?”–
Not possible.
The rate of input of information available to human brains, and the rate of integration into memory, and the breadth of paradigms in which information is structured, the precision of that information, synthesis of that information, and the volume of information in and across those paradigms exceeds by orders of magnitude the capacity of the human brain.
If you learn my (our) work on the unification of the fields and universal commensurability you can significantly improve the disambiguation, organization, synthesis and falsification of the available information. You’ll gain at least a standard deviation in demonstrated intelligence with the same IQ.
But assuming we either solve the algorithmic problem of the current models and their underlying hardware architecture, or we manage to build large numbers of neuromorphic computers, in three dimensional chips instead of two, and then using photons (light) solving the heat problem, then the necessity of pre-calculation, vast compute, and delayed integration of information, will be circumvented, and at that point leaving human capacity in the dust even more so than now.

Reply addressees: @RunicSigil

Q* or basically “recursive wayfinding” which all of us learn in first year of software enginering is simply adding (1) adversarial solutions (2) number of context anchors (3) wayfinding.

It’s what your brain does…. The problem is the speed and cost of reducing compute to accomplish it. I don’t see why this is going to be anything other than additional economic warfare over access to chips.

Reply addressees: @mark_k