Two things can both be true:
1) Prediction markets could be great machines to produce information, and can be monetized + valued as epistemic tools
2) To get to this point, they need to be much more widely distributed across people that can contribute valuable inputs - and the GTM to get there will depend on the information output. Selling PMs as generalized “forecasting platforms” is (and will continue to be) insufficient to achieve 1)!
"gambling/prediction markets like Polymarket"
Putting Polymarket into the category of "gambling" is a massive misunderstanding of what prediction markets are or why people (incl economists and policy intellectuals) are excited about them.
Prediction markets are interesting because they're a social epistemic tool: the public gets a view of how important certain events are and what kinds of things are likely to happen, that is much less vulnerable to biased editorial opinion than either social media or news websites. Conditional prediction markets have applications in governance, which we're starting to see already.