About

← Back home

This is an unofficial experiment in using AI for automated fact-checking of predictions. Podcasts are a fertile ground for predictions about the future, and we can create an automated pipeline that transcribes a podcast, extracts predictions, and validates whether or not those predictions ended up coming true via automated research on the web.

Here we evaluate the All-In podcast, which has a deep catalog of historical episodes and a set of hosts known for making predictions about the future. This repository could be forked and modified to evaluate other podcasts with minimal modifications.

Prediction classification descriptions

RightThe prediction appears to have been correct.
WrongThe prediction appears to have been incorrect.
AmbiguousIt's hard to tell if this was correct, due to confounding factors or unclear data.
InconclusiveNot enough time has passed to know whether this prediction was correct.