Before the next wave of AI agent boom arrives, ordinary retail investors need to "painfully change" their investment thinking. My personal suggestion is for reference only: 1) Fomo chasing the "AliExpress Flow" project: If the project goes online and is either heavily controlled by AliExpress's Dev or seized by a conspiracy group in the early stages, Fomo chasing the rise is likely to contribute short-term liquidity (rushing to the top of the mountain). Even if the project has sustained popularity and there are opportunities in the second and third stages, the psychological test and opportunity cost faced by most people are not proportional. It is better to wait for the next wave of early Alpha opportunities or find potential projects that have not yet been valued; 2) Purely based on Dev qualifications Github Repo、 Project narrative for investment: On chain investment projects do not have the endorsement of VC's "real money and silver" investment. Dev qualifications and Github repository standards have become the biggest reference benchmarks, but don't forget that Dev qualifications, Github star logos, etc. are easily mixed with moisture. If they are truly valuable, they can wait for the market Fomo to finish before getting on the car. Drawing lessons from the previous wave of major reshuffling, at least one high threshold condition for project build should be added to these value evaluation factors, after all, the starting technology and operational threshold for value projects must not be low; 3) The expected pricing and valuation are already locked in at 300-500 million yuan: it should be noted that the high chance of discovering Alpha on the chain is due to the removal of upstream exit pressure from VC and early participants. It is unreasonable to rush for projects over 50m or even 100-500 million yuan due to Fomo's leadership. Unless you believe that there is a target with a benign value support of 500 million yuan on the short-term chain, or you are confident that there is 3-5B room for this project. Obviously, the short-term track on the AI Agent chain is not mature at all, and only one @ aixbt_agent application scenario has been implemented, but further efforts are needed to increase the upper limit of commercial imagination. In short, do not bring the CEX coin selection valuation standard of 500M-5B from the previous cycle onto the chain; 4) Scattered and unfamiliar position allocation: The logic of position allocation is to concentrate chips as much as possible on valuable supporting currencies, so that there is confidence in surviving major drops and retracements until the rebound. If there is no research in hand and there are many asset targets in various PVPs, it is likely to suffer significant losses due to excessive junk assets, and the key is that it is likely to lose track confidence and be discouraged from being eliminated. Proper allocation and timely adjustment of warehouses may not lead to greater profits, but they can stabilize opportunities for long-term survival; 5) Always' pattern 'a certain currency: In the early stages of the development of the AI agent industry, the certainty is weak, and there are many asset targets. Value assets will inevitably be mixed with a large number of useless assets. If you accidentally target a certain junk asset, the diamond hand pattern will not only lose the principal, but also other better opportunities in the entire track. In the early stages of the track, it is definitely not wrong to use "scumbag thinking" in trading. Of course, it would be even better to continuously improve one's aesthetic and establish one's own core value currency position during the process. After all, being scumbag for a long time is not easy in PVP. Note： I know many people still want to recommend Ticker, but until the trend becomes clear, it is not possible to publicly recommend it on Twitter. I hope you understand. I suggest subscribing to my Substack column on the homepage to follow more content. Thank you all.