#预测市场 Seeing the data changes on the AI prediction market on Kalshi, I was suddenly reminded of the hype around prediction markets in the crypto space back in 2017. At that time, everyone was betting on various prediction platforms, confidently forecasting the future of different projects. Now, ten years later, very few projects that started out have survived and continued to iterate.

This time, Gemini's market share jumped from 30% at the beginning of the year to 86% today, reflecting what is happening behind the scenes in the real-time assessment of AI model performance. A trading volume exceeding ten million dollars indicates that participants are not just talking nonsense—they are voting with real money. Interestingly, around this time last year, ChatGPT was the undisputed leader, but in just a year, market expectations have flipped—such a dramatic change in probability reminds us that the competitive landscape in technology is far more fragile than we imagine.

Looking back at historical cycles, each time the front runner gets pushed back by the following wave. I have seen too many projects that once claimed to be "forever number one" fade away in the past. Prediction markets use objective leaderboards like LM Arena for settlement, which is at least much more honest than subjective public opinion. But this is also a trap—the market consensus can easily reverse suddenly at a critical point. With more than a month left until the end of the year, during this window, a major product launch or performance breakthrough could rewrite the probabilities.

Rather than chasing high recognition, it’s better to understand the underlying technological evolution. Gemini’s rise didn’t happen out of nowhere; it’s the result of years of accumulation and continuous investment by Google. Ultimately, this market will teach participants a simple truth: in a rapidly iterating field, past leadership can never guarantee future success.