This is quite a misleading talk, yeah include confounding variables with a fake dataset to make something look better. Tha fact is these methods have very little real world incremental value, most of the value goes towards resume embellishment. The data does not lie, you just have to look at it in the right way for your case with any method
While the video might not be particularly useful beyond pointing out the existence of and reasoning for bayesian causal inference, your comment is entirely useless. The data lies all the time. And what do you even mean by looking "at it in the right way for your case with any method"? Gibberish. Simulated data is being used to show that this approach is independent of problems that arise in a given research domain. It is not fake data. It is simply simulated.