The Web3 X AI Product Grid
As we kick off our deep dive series into NEAR Protocol at the intersection of Open Source and Decentralized AI, we want to start by clarifying the ‘landscape’ that the bulk of innovation for Web3 and AI will be happening. We call this the ‘Web3xAI’ product grid. It seeks to explain how blockchains can interact with closed-source, open-source, and decentralized-AI applications - and ultimately why blockchain is a fundamental part of AI’s development.
This is Web3 x AI Product grid:
Let’s break down the core of this grid: All AI innovation today is taking place across three domains. Closed Source AI, Open Source AI, and Decentralized AI. Ken Miyachi and Josh Daniels’ Web3 AI thesis from Lyrik Ventures, does a nice summary on their differences:
This ‘core’ of the Web3-AI product grid, is the foundation for where any type of AI application or Large Language Model will exist from. Any or all of these types of models have the capacity to plug into blockchain in two fundamental ways: 1) The payments layer, 2) The governance layer.
The Payments Layer
The top of the Web3 AI product grid focuses on payments. One of the lowest hanging use-cases for Web3 AI is AI-Agent payments on blockchain - simply because an AI will never be able to have a bank account or pay KYC on its own. NEAR Co-Founder, Illia Polosukhin speaks about this in the clip below:
The Payment layer of the grid, simply represents the potential to integrate token or crypto-currency payments for the monetization of an AI application in question. This can be done for closed-source AI companies (although it is not usually), open-source AI access, and by necessity, for decentralized AI.
The Governance Layer
The bottom layer of the Web3 AI product grid focuses on governance of the AI system in question. While closed source AI is managed privately by a foundation or a company, the future of AI does not have to continue in such a direction: Multi-Sigs and DAOs enable shared governance over the parameters and decisions surrounding all types of AI models: Closed, Open, and Decentralized. This governance may range from defining who can train the AI, to who can vote on updates or access, to who can shut an AI down. In some scenarios, an LLM itself could help govern another AI system!
Illia speaking here on AI and governance with Vitalik:
On the whole, this grid should be a reference to help orient you towards the future types of innovation we see emerging at the intersection of Web3 and AI. Future articles focus on the value proposition of how on this grid, a new world for AI transparency, payments, safety, and network effects is originating.
