Author: Li Jin, Partner at Variant Fund; Translated by 0xjs@ Jinse Finance
I spend a lot of time researching the intersection of AI and cryptocurrency. I believe that the combination of the two is critical to making AI more fair and user-owned.
Just sharing some quick thoughts on what we saw:
Projects at the intersection of AI x cryptocurrency fall into two broad categories: 1) Applying crypto to solve AI problems 2) Applying AI to solve crypto problems.
Applying encryption to AI
Since the AI market overall is larger than the crypto world today, most companies are focused on (1) applying AI to crypto. That is, applying AI to cryptocurrency (2) can make cryptocurrency easier to use and interact with, thus helping to expand the pie.
In the first category, we are looking at projects across the entire AI stack:
Data collection (including collecting new proprietary datasets that do not exist elsewhere)
Model training and inference (including data labeling and RLHF)
Infrastructure Market (including computing and storage)
AI Applications
We see opportunities for cryptocurrencies to have an impact in all of these categories.
Taking the data collection layer as an example, what we see is that amassing datasets that currently do not exist or are difficult to access could be a huge unlock for AI companies. Cryptocurrencies can help by incentivizing users to contribute data (as in many DEPIN networks).
In the compute market, we are also seeing token incentives being used to aggregate the supply side of these networks. We believe that in an increasingly competitive environment, a differentiator will be offering additional services and models to attract developers.
At the application layer, we see projects that aim to create AI agents, equipped with wallets to enable them to conduct transactions independently, or AI agents that run on decentralized computing (like Venice.ai).
We’ve made several early investments at the intersection of crypto + AI — Worldcoin, Mana, Modulus, Botto, Dawn Wallet — and we’re excited to continue supporting the space.
Applying AI to Crypto
Below I’ll outline the second category: companies applying AI to cryptocurrency.
The application of AI in crypto is similar to the broader impact we’ve seen AI and big language models have on non-crypto: creating a new computing user interface; aiding content creators in their creative process; and helping businesses improve their operational capabilities.
The categories of projects applying AI to crypto can be segmented by GTM: consumer-oriented, creator-oriented, or developer-oriented.
On the consumer side, the project is building:
AI interfaces for users (e.g. natural language interfaces for wallets to process transactions on-chain)
AI Trading Agents and Strategies
Valuation models (e.g. NFT Upshot)
User-owned AI DAOs (e.g. @hollyplusp_voice model, @BottoDAO)
On-chain AI games (e.g. Leela from ModulusLabs)
For creators, there are AI-assisted creation tools (such as title.xyz for NFTs) as well as AI-generated memecoins, and AI art in particular is becoming popular in the NFT world.
For developers, projects are applying AI to: Smart contract auditing; Fraud detection and security; On-chain data analysis; Community management via chatbots; Defi risk scoring; Sybil detection; Token distribution.
