using a byproduct of mold to cure a bacterial infection isn't simple. Merely noticing it was a rare occurrence (Fleming got lucky with his dirty lab), it only

using a byproduct of mold to cure a bacterial infection isn't simple. Merely noticing it was a rare occurrence (Fleming got lucky with his dirty lab), it only works on gram positive bacteria, and importantly, use of penicillin leads to evolution of antibiotic resistance, reducing the utility of penicillin. That last point especially is a highly non-simple process. Like just about anything in medical biology, there is a simple narrative taught to the general public, and then there's the reality as it stands in modern medical research and practice. For prions, I'm sure people can find any number of ways to module prion activity in vitro, most of which won't work in vivo/in the real world, for messy complex reasons. To use machine learning parlance, a feature like the existence of a TED talk by the creators of a medical treatment being True highly correlates with the ineffectiveness of the treatment (same as "If published in Nature, it's almost certainly wrong!")