My ideal AI+blockchain project (3)

Source: Daoshuo Blockchain

- Story Protocol: https://www.storyprotocol.xyz /

The problem this project focuses on is how to enable content creators to maximize their own interests in the era of encryption, while on the other hand allowing the public to build on existing new innovations in content.

The way it solves this problem is to define the content (that is, IP Assets, IP assets) as "Nouns" (nouns), and the way to process the content Defined as "Modules" (modules) or as "verbs" that operate on nouns.

Content can be represented by NFT on the chain, and the way to process the content can be defined by various methods (smart contracts).

Under this framework, a series of issues such as how to protect intellectual property rights and how to protect the interests of content creators can be developed in the definition of Nouns and Modules. The author is free to play.

This agreement actually revolves around data and data processing, redefining data and designing a framework for data processing methods. In my opinion, it is more like using blockchain technology to solve some problems in the field of big data, and it is not so closely related to AI.

Of course, more people will still say that data is also a part of AI, and this project is certainly considered AI.

In addition, it is conceivable to add AI processing methods to Modules in the future to improve the efficiency of data processing.

So from these perspectives, it is not an exaggeration to say that this project is AI+blockchain.

I think this kind of innovative project from an architectural perspective is quite interesting.

But the uncertainty of this type of project is that it creates a structure, but how to use this structure specifically and whether it can achieve the expected effect still requires a lot of practice. Testing and exploring, I’m afraid we won’t see real results in the short term.

I classify this project as using blockchain to solve the incentive problem of AI data creators.

- Space and Time: https://www.spaceandtime.io/

This The vision proposed by the project is very great. It is to build a decentralized data warehouse

It gives a specific usage scenario: linking the chain The on-chain data and off-chain data are integrated and then sent to the smart contract.

In this process, an obvious problem is that the data may be tampered with during integration or processing. So on the one hand, this project indexes (all) the on-chain data of the blockchain, on the other hand, it overlays the off-chain data, and then uses zero-knowledge proof to ensure that these results have not been tampered with before sending them to the smart contract.

Because it strives to ensure that data cannot be tampered with, it also defines itself as a "verifiable compute layer".

From the introduction of the project, I think it does more to prevent data tampering, which is not the same as the "decentralized data warehouse" I imagined. Too similar.

In addition, logically there is no problem with the idea of ​​"anti-data tampering", but there is a question about the specific solution of this project:

In this process, using zero-knowledge proof to process data is a key link in the entire process. Whether this link is decentralized is the focus of my attention:

Is it a decentralized network layer or a centralized operation monopolized by this project? ?

If it is a centralized operation monopolized by this project, then it is the weakest point in the entire system.

In my opinion, this project actually uses blockchain (encryption technology) to deal with problems in big data, more specifically using blocks Chain technology solves the problem of anti-tampering in AI data.

Reference link:

"Quick Overview of 11 Mainstream Encrypted AI Projects":
https://www.jinse.cn/blockchain/3677551.html