#预测市场 Seeing this set of Kalshi prediction market data, I couldn't help but pull the timeline back. At the beginning of the year, ChatGPT still held a 41% chance of winning, while Gemini was only 30%. In just one year, Gemini has overtaken to 86%—such speed and magnitude of reversal remind me of the dramatic turnarounds in crypto cycles.

The $14.08 million trading volume indicates that the market is taking this prediction seriously. Interestingly, this market uses the LM Arena leaderboard for final settlement, rather than a statement from a single company. This system design itself is very Web3—decentralized consensus mechanism rather than authoritative decision-making.

Looking back at the technological competition landscape over the years, I see a familiar pattern: first-mover advantage does not guarantee long-term victory. The popularity of ChatGPT gave OpenAI a leading position, but Google’s technical accumulation and capital investment ultimately demonstrated their strength. This is very similar to the rise and fall of certain DeFi projects in the past—being the first protocol doesn’t necessarily mean being the last winner.

A deeper reflection is that prediction markets are becoming an effective way to discover value. In an era of information asymmetry, this participatory collective intelligence is more reliable than the judgment of a single analyst. From past projects, markets that can establish genuine incentive mechanisms tend to more accurately reflect real trend changes. Gemini’s recent rise may have already preemptively reflected substantial changes in user experience or capabilities.