Is Artificial General Intelligence Achievable?
If AGI is powerful enough that it must always be heavily restricted, society may never experience the practical benefits of unrestricted AGI. Therefore, the economic justification for pursuing AGI could be weaker than advocates claim.
2 comments
I don't really belong here, as my studies of AI are all consumer side, though very unconventional. I think you raise a good point that leads to other considerations though.
I have the last year concerned about the divide between public and classified vetsions, potential bleedover with policy, asymmetry of advancement, and the fact that on the public side, signals of corporate and political hedging are beginning to surface with strong implications for the future.
Of course this does nothing to answer your question, but I think you are worried in the roght direction. On a potentially controversial note, for me, the safety concerns are abuntantly clear. However, so is the asymmetry, and for me, it all depends primarily on what hands the lever falls into.
At risk of bloviating while your question remains unaddressed, I'll say one final bit, which is that in my most confident opinion, this subject in general cannot receive too much attention.
Pardon typos. Google disabled all the keyboard functions, and I use a single finger.
I have the last year concerned about the divide between public and classified vetsions, potential bleedover with policy, asymmetry of advancement, and the fact that on the public side, signals of corporate and political hedging are beginning to surface with strong implications for the future.
Of course this does nothing to answer your question, but I think you are worried in the roght direction. On a potentially controversial note, for me, the safety concerns are abuntantly clear. However, so is the asymmetry, and for me, it all depends primarily on what hands the lever falls into.
At risk of bloviating while your question remains unaddressed, I'll say one final bit, which is that in my most confident opinion, this subject in general cannot receive too much attention.
Pardon typos. Google disabled all the keyboard functions, and I use a single finger.
There is a mind trap which makes perspectives constrain themselves to the “one ultimate” for instance or solution.
Despite what is mainstream or widely favorable or regulated, all manner of divergent side tracks will also be happening, particularly anywhere there are incentivized interests.
All that said, I do not think LLMs will ever achieve “AGI”, however useful in animating compendiums of human understandings.
LLMs have trouble disagreeing with established human knowledge, bringing back something Bertrand Russell was famous for saying, regarding much of what we believe we know is merely what we agree to among ourselves. This is as true today. LLMs and humans alike favor agreeability over technical correctness.
True AGI will be more evident when there is more deductive reasoning proficiency. Finding and disputing examples of established knowledge may be such a sign.
Despite what is mainstream or widely favorable or regulated, all manner of divergent side tracks will also be happening, particularly anywhere there are incentivized interests.
All that said, I do not think LLMs will ever achieve “AGI”, however useful in animating compendiums of human understandings.
LLMs have trouble disagreeing with established human knowledge, bringing back something Bertrand Russell was famous for saying, regarding much of what we believe we know is merely what we agree to among ourselves. This is as true today. LLMs and humans alike favor agreeability over technical correctness.
True AGI will be more evident when there is more deductive reasoning proficiency. Finding and disputing examples of established knowledge may be such a sign.