I was shocked when my market on Amy Klobuchar's claim that the TAKE IT DOWN Act would pass resolved YES.
It was a prediction on a podcast. It was a prediction on a podcast by an elected official. It was a prediction on a podcast by an elected official in the minority party. It was a prediction on a podcast by an elected official in the minority party about internet/AI regulation. It was a prediction on a podcast by an elected official in the minority party about internet/AI regulation that would resolve in a single-digit number of months. Those never actually happen, right?
But the specificity of the claim itself should have been some signal in favor of the claim. Not because the specifics of the claim made it more likely, but because the specifics allowed it to resolve AT ALL. Most claims on podcasts are outright bullshit and, if you did make a market out of them, the resolution criteria would be ambiguous at best.
I've wanted to make more markets about claims on podcasts but this is the problem I keep running into. I don't want my markets to be all judgement calls on ambiguous resolution criteria. And no, I don't want to farm this out to LLMs to cook up plausible, objective, but ultimately slopified resolution criteria. My pipe dream is that people would do this work for me when they speak publicly. In that regard, thank you to Senator Klobuchar.