polynya on Warpcast

polynya pfp

Though well-written and inspirational, I believe this post fundamentally misunderstands Ethereum and blockchains. Ethereum's distinctive property is objective, strict global consensus, and absolutely nothing else. It's great for usecases that require strict global consensus, but it's impossible for everything else, because Ethereum cannot parse any subjectivity or rough consensus at all. As such, while money, contracts, governance, identity, law are presented as examples, Ethereum can only parse very, very limited forms of the above, where it's objective. In some cases, like governance or law, it's almost entirely subjective with negligible scope for Ethereum to help. 99.99% of economics, institutions and the like are deeply human and subjective, which Ethereum or blockchains in general cannot interpret at all. Indeed, we've seen many a times how forcing subjectivity into objective code has led to many disastrous outcomes in crypto. (Contd...)

7 replies

12 recasts

96 reactions

polynya pfp

To be clear, Ethereum is extremely valuable in usecases where strict global consensus is required - autonomous store-of-value, "dumb" contracts like basic DeFi, objective identity registries, but it's also important to understand this so that we can focus on these usecases where it makes a difference, and not waste time where it cannot. We have wasted billions of dollars and countless person-hours on these diversions because of a poor understanding of strict global consensus.

5 replies

3 recasts

44 reactions

Sylve pfp

How do you picture dumb contracts interacting with the valuable subjective world? I believe zk-ifying oracles is one way to do it, i wonder if you're looking at anything else

1 reply

0 recast

1 reaction

polynya pfp

Haven't really thought about it, kind of just thought oracles are the interface between the objective and subjective worlds, and yes, improving by ZKfying them are the way to go I guess maybe a long way down the line, as AI develops to be more general, it can replace the whole blockchain+oracles apparatus to be directly more subjective

2 replies

0 recast

2 reactions

Sylve pfp

I like ideas like semantic zkTLS for https://tmr.news/: use an external LLM to give a subjective idea of "how close you were" That doesn't have to be zkified imo, you can just use signatures for it

0 reply

0 recast

1 reaction

Davide Crapis pfp

I don't think that as AI develops it will replace blockchains+oracles, on the contrary there will be many AIs and they will be in need of making hard objective commitments to one another

0 reply

1 recast

1 reaction