[ExI] Limiting factor to the Intelligence Singularity?

[Keith Henson]

On Sat, Dec 23, 2023 at 8:40 AM Jason Resch via extropy-chat
<extropy-chat at lists.extropy.org> wrote:
>
> On Sat, Dec 23, 2023, 11:12 AM Kelly Anderson via extropy-chat <extropy-chat at lists.extropy.org> wrote:
>>
>> Einstein proposed that nothing could achieve the speed of light
>> because it gained mass as it got closer to that speed.
>>
>> I wonder if something similar might affect intellectual progress.
>
>
> There is. Seth Lloyd describes the physical limits of computational speed and memory density, which are a function of just three physical constants (c, h-bar, and G): https://arxiv.org/abs/quant-ph/9908043

I have to read this more carefully, but I wonder if he got the
radiation limits right.

If Tabby's Star blinking is the result of huge data centers coming
between us and the star, then we have an example

Kieth

> I would argue these computational bounds imply a physical bound on intelligence (which is limited by, and to an extent a function of, computation and memory).
>
> Roughly speaking, computational speed is a function of mass, while memory is a function of mass and volume. The greater the volume the more memory, but the slower non-parallelizable processing becomes. Energy use is a function of forgetting (erasing information) and background temperature. A reversible computer loses no information, and therefore need use no energy at all to run.
>
> The physical object with the fastest computation and memory density would look like a black hole:
> https://www.scientificamerican.com/article/black-hole-computers-2007-04/
>
>
>> Though there is no c there, so it might not hold together entirely.
>> Hang with this for a second though, as there is a counter exponential
>> at play.
>>
>> As we grope into the future, there is a trend away from the single
>> brilliant inventor. Edison had a whole lab of people, some of whom
>> were clearly smarter than he. Eli Whitney pretty much invented the
>> cotton gin by himself relying on only a few simple precedent ideas. By
>> the time we get to super colliders, fusion reactors and rockets, you
>> need a village of scientists and engineers to make progress.
>
>
> I see echos of the memory / parallelism/ volume / speed trade-offs in physics reemerging here in the context of human brains, teams, remote work, etc. Very interesting.
>
>>
>> Is it possible that the exponential curve towards the singularity has
>> a hidden negative signal of increased resistance to progress because
>> of the required size of the team? Might this be one reason that we
>> haven't yet progressed beyond the jumbo jet airliner? Is that
>> indicative of the future of a lot of things? How many Intel employees
>> does it take to design the next iteration of their CPU? Yes, they have
>> computational minds that help them with layout and such now... but you
>> can hardly say Oppenheimer was the inventor of the atomic bomb. He was
>> the face of the project and coordinated the efforts to some degree,
>> but you can't make that level of progress without a sea of minds. Musk
>> has ideas (or maybe his people do that too) and finances them, but the
>> actual work is carried forward by an army of engineering ants. As I
>> believe in emergence as a deep concept, I tend to see groups more than
>> individuals, though I value individual contribution greatly.
>
>
> Progress has loosely been a function of the number of inventors which is a function of population. There was a trend change in population growth in the 60s which seemed to alter the path we were on:
>
> https://slatestarcodex.com/2019/04/22/1960-the-year-the-singularity-was-cancelled/
>
> I don't know to what extent this may have forestalled thy singularity, however, as computers have been taking on an ever greater fraction of the cognitive load. I think we're still roughly on track for van Forrester's predictions of 2027 +/- 5 years.
>
>
>>
>> Perhaps my mind has wandered too freely and it is time to go work on
>> my own, much simpler, inventions, as they seem to constantly be
>> broken.
>
>
> Thanks for your intriguing ideas. I appreciate them.
>
> Jason
> _______________________________________________
> extropy-chat mailing list
> extropy-chat at lists.extropy.org
> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat