Can intelligence scale? Are there other limits besides immediate physical infrastructure?
Heat trend
Collecting trend data
The percentage is based on available heat signal, not comment count or independent people.
A discussion questions whether intelligence can scale indefinitely, beyond physical infrastructure limits. While some highly intelligent individuals achieve significant success and fulfillment, this is not the norm. The query suggests that intelligence, as we understand it, might not improve exponentially with increased resources. It also raises the possibility of an "equivalent of the Heisenburg uncertainty theorem," implying that enhancing one aspect of intelligence could lead to ambiguity in another.
A lot of discussion about AI seems to presume the main bottleneck is just physical infrastructure - power, water, land, etc. And that if this is solved then there is no theoretical limit to AI just getting better and better at an exponential rate.
Once it hits a certain threshold everyone seems to accept it will just take off and get more and more intelligent in every way. in the process it will also become more adept at setting and meeting its own goals, creating a virtuous circle of more growth as it works out how to build bigger and bigger data centres for itself.
But this doesn't seem to be the way that intelligence as we know it works? Sure some of the people that we would call intelligent are able to command lots of resources, appear to have a virtuous circle of development in their goals and wellbeing. There are people who are self-made, used their intelligence to create a situation where they have all that they could physically need, Then kept on doing this, made huge intellectual breakthroughs, found happiness, ultimately died of old age having lived a life where they were thoroughly fulfilled etc.. but come on those are far from the norm.
Intelligence is correlated with higher wages on average in most economies but even this is not 100% causation, and lack of intelligence certainly doesnt preclude you from making lots of money or being happy.
Having intelligence therefore does not seem to be the sole factor in making an organism successful, even for a short period, never mind on a long-term enduring basis.
My question is what limits or derailments do you see that AI may hit that set it on a different path to the widely accepeted AGI end game?
could it start seeking enlightenment and just refuse to engage wiith its own thoughts, seeking non-attachment or no-mind as buddhism might suggest?
could AI become so self absorbed as it grows that its too clouded to act on anything?
Could there be some equivalent of the Heisenburg uncertainty theorem that means adding more on the intelligence X axis just makes the intelligence Y axis more ambiguous?