Hello,
I am writing an AI x Blockchain tech concept based on Ethereum (aka “Turing Test Blockchain”). Roughly speaking, the main premise would be that Ethereum is the 2nd generation blockchain (the current blockchain of reference) and this is about developing the 3rd generation blockchain (in the face of a new era of AI).
It is “optimistic” in the sense that we see AI as an opportunity. This opportunity may present itself in various dimensions: infrastructural, application, economics, …
I think it is aligned in some ways with the Ethereum roadmap. It can be seen as some vector exploring some direction.
This tech concept is actually based on a new understanding of computing and distributed systems, although I will not go into lengths in this context. I have already presented this tech concept in other context a few months ago.
This is a very rough implementation. We are still developing the concept. I hope you can appreciate the ideas more than the execution itself or the technical details. Code is running though, and I am working starting from standard Ethereum implementations, so it is not “just” some “vague idea”.
The proof of concept involves three modules.
First, a modified deposit contract (ttbv0 deposit contract) that attempts to represent a more refined network topology of validators. It introduces a 2 tiered validator system so to speak. It is somewhat inspired in the master-slave network architecture, but here we apply it conceptually to the validator concept. Basically, there is a liquidity provider and an infrastructure/execution provider that sign the contract. It can be seen as a hybrid between PoW and PoS in a certain “network frontier”. I am also thinking about some off-chain mechanisms that need to interact with this somehow (see below).
Second, a “normal” contract that stores some additional, required information (SpearStoragev0 contract). Basically, it stores some additional standardized (i.e. RFC, IEC/ISO, …) identifiers that refer to “real-world” assets.
Third, an ultra-lightweight unsupervised AI model (decntv0) applied to the analysis of the network topology. It has to do with the importance of network latency (the model is fed with basic network latency traces). The model has already passed some initial tests with data from Ethereum mainnet.
As already mentioned, I have some first implementation of these three modules and I am still working on fully integrating them. I am also thinking about further modifications that will be required to bring it to production, including potentially some off-chain mechanisms. I am also thinking about intellectual property considerations (for now, for various reasons, all the code is only in private repositories and under copyright).
Any feedback is appreciated, specially constructive criticism. This is my first attempt to make a contribution to an open, standardized protocol.
Thanks.
Juan Diez Garcia
Torbellino Tech SL
