The Bayesian in-joke: everything that works in ML is 'approximately Bayesian' anyway

burny_tech · x · 2026-11-19

A classic ML community quip: nearly every method that ends up working in machine learning can, in hindsight, be framed as some approximation of Bayesian inference — a long-running joke that reflects the framework's catch-all status in ML theory.

Original post →

More from Fun

Fun channel →