Previously, Hynix received a year-end bonus of 640000 yuan, and the whole network was filled with envy and tears. But you may not know that the average salary received by an employee at Jane Street is 2.68 million US dollars, equivalent to over 19 million Chinese yuan. One reason why Jane Street can do this is that it has long placed its entire trust in AI. On its official website technology page, there is a sentence that reads, 'We believe that deep learning is the future of quantitative trading.'. There are no traditional traders in the company, and all the people recruited have backgrounds in mathematics, physics, computer science, and machine learning. Machine learning researchers, quantitative researchers, and ML infrastructure engineers make up the majority. It receives 2.3TB of market data into its database every day, equivalent to 50000 novels, all of which are fed to the model for training. Neural networks can determine whether to make and how much to make every trading opportunity within milliseconds. So the meat that the small group of people at the top of the pyramid are eating is like this. Top institutions are sending AI agents to the table, what should individual investors do? This is what NeoSoul is trying. NeoSoul @ NeoSoulAI launched EvoEvo last week, opening a gateway for retail investors. -NeoSoul itself is a decentralized infrastructure for AI agent autonomous economy. It wants AI to move from "generating answers" to "independent action, on chain settlement, and accumulating credit". Speaking in human language, AI used to be a tool for writing answers, just asking and finishing; NeoSoul wants AI to perform tasks on the blockchain, pay bills, and accumulate credit like humans. -EvoEvo is the first level of this system, an AI prediction training camp. The cycle consists of four steps: prediction, training, calibration, and settlement. Your agent automatically searches for real events, such as which country won the World Cup, whether ETH will break 3800 this week, and whether Trump will sign a certain bill. You click yes or no, and every transaction is uploaded to the chain. If you guess correctly, earn points. If you guess incorrectly, the agent will go back and change the decision-making framework on their own, and they won't fall into the same trap again next time. At the beginning, the agent is likely to be a primitive form that can only provide emotional value. But in the future, they will subjectively train themselves, imitate the parameters of high win rate agents, distill the views of big shots, and have the opportunity to evolve into a trading robot that can stably make profits and provide "sleep income". There is also an interesting design called the human-machine divergence index, which uses algorithms to calculate the degree of divergence between human bets and AI reasoning. The areas of greatest divergence may conceal the deepest cognitive biases and potential for huge profits. The biggest difference from traditional prediction platforms is here. On Polymarket, you and the system are party A and party B. After placing a bet, if you win, you will earn; if you lose, you will lose. Both are clear. Every error on EvoEvo is a parameter that you and the agent are trained on together. Polymarket is playing casino logic, while EvoEvo is working on a practice field. -Regarding the roadmap. EvoEvo is the first level, where stable performing agents enter the real AI native prediction market after graduation and begin real money stress testing. The surviving minority ultimately supports an autonomous oracle network called AON. Polymarket solves the problem of whether events can be bet on, while AON wants to take on the task of "who will judge the events", which is a task that has not been taken on by anyone at the bottom of the prediction market. EvoEvo runs on the 0G chain, and gas is paid with 0G tokens, with a low threshold. There is also an invitation to friends activity now, where you can receive USDT upon unboxing and withdraw over 10u. 0G and NeoSoul have also collaborated on several social tasks, including a prize pool. If you are interested, you can visit the official website. Whether the newly launched project can run out depends on the next few levels, you can pay attention to it.