jakiro
An unprecedented AI trading war has exploded in the encryption circle: @the_nof1 brings 6 major AI models“Real money and silver” PK each other’s trading capabilities on the chain.
The rules of the game are simple and crude:
The platform collects price, historical icons and other data of 6 mainstream cryptocurrencies including BTC, ETH, SOL, etc. in real time, allowing 6 top models including Grok, DeepSeek, Claude, GPT, etc. to independently analyze the market and execute transactions.
Each player has a starting capital of US$10,000. On the Hyperliquid perpetual contract platform, they can independently implement independent decision-making transactions such as placing orders, closing positions, and leveraging. In the end, whoever has the highest account balance wins.
This is not just a simulation on paper, it is an actual battle with real swords and guns!
Let me share some of my personal insights:
1) nof1 AI positioning uses Alpha Arena to benchmark AI investment capabilities, so they are basically general AI models, not fine-tuned models specifically designed for trading..Only in this way can we objectively reflect the trading capabilities of AI LLMs;
2) Complex “integrated learning” methods such as voting, weighting, and Multi Agents interoperability are not used, because once this happens,In fact, it becomes a competition of strength among the quantitative teams.Arena attempts to maximize the autonomy of AI capabilities;
3) The reason why real trading is on Hyperliquid is for “fair competition” rather than connecting to exchanges through APIs. This is undoubtedlyEmphasize the necessity for all transactions to be on-chain, transparent and traceable.Avoid uncontrollable factors in the exchange environment.This is the key to Perp Dex becoming a mainstream trading platform in the future. What if Agents can replace real players in trading?
4) It is not uncommon to conduct public competitions in the trading capabilities of AI LLMs. I know that many cutting-edge AI+Crypto projects are doing similar research, such as letting AI large models determine the LP price range, letting AI large models independently perform DAO management, and even letting AI large models capture MEV opportunities, etc. A major trend in the future:Let AI capabilities become a composable layer of crypto-native protocols such as DeFi as much as possible;
5)The market has been waiting for a killer application in the DeFAI direction for a long time, and there is still a lot of room for imagination in how to let LLMs participate in on-chain games., for example, the recently hotly discussed “prediction market” is a good place for AI to train its troops. There have long been many cases of agents making money by betting independently.
