I personally believe that there may be three ways to predict the existence of liquidity in the future market 1. CLOB - Polymarket/Kalshi - Player opponents mainly market makers - suitable for large markets 2. Bonding Curve - 42， There is another prediction market I am looking at - players against other players on the curve - suitable for localization+long tail market 3. HLP+AI - It is uncertain if there are any projects that do this, and the funding source can be a HLP model similar to Hyperliquid. Then, the market maker (AI) automatically judges the likelihood of this event every time a new market is opened, and how much percentage of funds should be allocated to the initial liquidity pool. The form can be either LMSR AMM used by Polymarket before or CLOB - theoretically, both large markets and long tails can do this, but obviously, long tails without real person market making are more suitable In this case, the player's opponent is an AI market maker, which is different from human market makers who ignore probability and only engage in price arbitrage. AI market makers also engage in "AI judgment and human probability judgment gap" trading arbitrage - of course, this premise is that LLM's judgment of most markets is better than that of humans, in order to ensure that "HLP" can make a steady profit To solve the problem of opponent's plate, the earliest SNX's unified flow pool+Leviathan mode with all players' PK win rate, and later Overlay's Rebase mode with reference to BAS, were also based on the idea of "automatically finding opponent's plate". However, there are few projects that should be used now