Decentralized Identity Trilemma(maciek.blog)
maciek.blog
Decentralized Identity Trilemma
http://maciek.blog/decentralized-identity-trilemma/
4 comments
Some more discussion on the topic over here: https://bitcointalk.org/index.php?topic=4459113.0
Interesting idea, thanks for the link.
I think the direction is right -- the solution probably lies in deriving scarcity from intelligence unique to the brain (for now).
Both brains and computers can perform pattern recognition. Computers are better at this task because they can be scaled and brains cannot. Therefore, this form of intelligence lies in the domain of financial capital. It can be abstracted through the cost of performing arithmetic operations (Proof-of-Work).
Brains can perform abstract thinking (answering the question why), computers cannot. If we could create a metric tied to the ability, it would be nested in the human capital.
I'm skeptical in regards to the practical application proposed in the thread. Even if this is robust, it'd be incredibly wasteful. The whole point of PoW is that it's moving computation from the brain onto a chip. This would go the other way around.
I think the direction is right -- the solution probably lies in deriving scarcity from intelligence unique to the brain (for now).
Both brains and computers can perform pattern recognition. Computers are better at this task because they can be scaled and brains cannot. Therefore, this form of intelligence lies in the domain of financial capital. It can be abstracted through the cost of performing arithmetic operations (Proof-of-Work).
Brains can perform abstract thinking (answering the question why), computers cannot. If we could create a metric tied to the ability, it would be nested in the human capital.
I'm skeptical in regards to the practical application proposed in the thread. Even if this is robust, it'd be incredibly wasteful. The whole point of PoW is that it's moving computation from the brain onto a chip. This would go the other way around.
That idea is oversimplified. One can write a traveling salesman solver, that will outperform most humans in several weeks of work: any decent RL algo like A2C will do. Even a simple genetic algorithm will probably perform better actually, and that can be written as an exercise during a CS school lesson.