Using LLM on top Prediction Markets to understand the coming news ?

I’ve been experimenting with w way to pull together prediction markets signals with LLM to create a narrative around the data. It is easier for me to read and remember a story than a stack of data. I have tried on crypto, stocks, geopolitics and politics—and now thinking to package them into straight to the point Substack updates. The same data point can be spun ten different ways (bullish, bearish, risk-on, risk-off…), so I’d love your take. I put an example in the link.