I'm sure the use of prediction markets to predict existential threats is difficult, but it seems like you could at least use them to predict the emergence of AI. I'd be surprised if this wasn't discussed here at some point.

It seems to me that while prediction markets may not need funding from a technical perspective, public and especially political opinion on them does need some nudging. I don't think I'm entering mind-killing territory by suggesting it'd be good if politics didn't get in their way so much. I'm certainly no expert, but long-running markets where investors would expect interest paid would face all sorts of (US) regulatory hurdles beyond normal markets. Its very expensive just to find out what regulations you'll run afoul of, as obviously US financial regulation was not created with prediction markets which could last decades in mind (or prediction markets at all for that matter).