#预测市场 Seeing the AI prediction market data on Kalshi, I have to say honestly—stories behind these probability reversals are more worth caution than the numbers themselves.

At the beginning of the year, Gemini's winning probability was only 30%, now it has skyrocketed to 86%. ChatGPT's probability dropped from 41% to 8%. $14.08 million in trading volume sounds lively, but you have to ask yourself: is this market rational judgment, or another round of expectation gaming driven by capital?

I've seen this too many times in this circle— a hot probability surges all the way up, retail investors follow suit and go ALL IN, only to find out at the time of settlement according to the LM Arena leaderboard that the market's price discovery ability isn't that magical. The problem is, by the time of settlement, it's already too late—someone has long since withdrawn.

These prediction markets seem quite genuine, but don't forget a few risk points: First, the issue of liquidity concentration—big players can easily influence the probability; second, the settlement standards themselves have gray areas, and LM Arena's rankings are not entirely objective; third, the participant structure—if mainly project teams or stakeholders are trading, then the probability is just their chip game.

What should genuine investors do? Don't be fooled by data from a single prediction market. Look at who the counterparties are behind the trades, check if the liquidity depth is sufficient, and most importantly, understand that high probability in a prediction market ≠ the event will definitely happen— it only reflects the current consensus of market participants, and that consensus is often changeable.

This round of Gemini's probability surge may only reflect a temporary trend, not the final outcome.