Even if they actually did work as well as the author wants to believe they could, there's still the matter of utility. Efficient pricing of real assets actually injects something, greases the gears of the market.

What is the utility of knowing the odds of something? Well that depends entirely on the something - it needs to be actionable, affect something that is not merely itself, whereby it can lessen friction. But this "actually valuable?" function is entirely decoupled from the prediction market trade volume and odds! You can make bank contributing to entirely useless odds, and equally make little on something where a clear prediction actually solves problems.

Those who stand to benefit from the output information (if ignoring insider bullshit) don't have to play. Hell if it's something important enough and they really have an edge, why would they play at all, if the information or prediction truly is valuable?

VCs don't exactly make plays on merely doubling their investment, right? Plus who's to say those sitting on valuable modeling powers actually have the capital to meaningfully participate? Even literally knowing the future wouldn't mean much if you're always only betting fifty bucks and spending the proceeds on heroin or tokens or whatever else is fashionable these days.