Something terrible is happening Thinking before bedtime, who will be the ultimate winner in the era of AI trading? If future transactions are fully AI driven, the essence of the market will evolve into a stock killing of computing power and logic The appearance of trading is a game of funds, and the core is a competition of human nature. AI is essentially a logical collapse of historical data, capable of approaching rationality to the extreme, but unable to quantify the unpredictable irrational flicker in human nature When all traders wear the mask of AI, the homogenization of algorithms will inevitably lead to strategy crowding and capacity bottlenecks. At that time, it was no longer the people who executed logic who made money, but the very few who could 'predict AI's predictions' or search for human warmth in algorithm dead ends Further reflection: ⬇️ Who would lose money? (Sacrifice of algorithm homogenization) Moderate AI users: When everyone uses similar open-source models or mainstream strategies, AI will generate convergent trading. Everyone buys and sells in the same millisecond. The result is that profits are consumed by extremely high friction costs (slippage), or collectively liquidated in violent fluctuations Dogmatists who rely on historical experience: AI is an inductive method based on historical data. Once an unprecedented "black swan" event occurs in the real world (such as unprecedented geopolitical conflicts), AI will fall into collective paralysis due to logical failure. 2. Who will make money? (The controller of dimensionality reduction attacks) Rule makers (computing power and channel giants): Regardless of who wins or loses, exchanges and hardware vendors providing fast computing power will always receive a commission. This is the so-called 'shovel seller' Metastrategy developer: someone who can identify AI logic vulnerabilities in the opponent's disk. If I knew that your AI would definitely stop losses when it fell below 5%, then I could accurately use your stop loss order to complete my position Embrace those who are 'extremely irrational': art, intuition, and taking a gamble. These things that are considered "noise" in the model are precisely the most difficult areas for AI to understand Why does AI face a 'capacity bottleneck'? Liquidity depletion: Trading requires counterparty positions. If the results calculated by AI are all 'should buy', then there will be no sellers in the market. The consistency of this logic can cause the market to instantly lock up and lose liquidity The disappearance of mean regression: AI is too fast, it will erase all profit margins within one percent of a second. In the end, the market may enter a state of 'dead silence' unless there is new information injected that cannot be understood by AI Final summary AI can beat ordinary people, but it is difficult for AI to beat the 'all AI market'. In that era, 'pure rationality' would become cheap, and 'high-quality irrationality' would instead become a luxury item