> Consider the impact that simply believing the relevant portions of the world has been modeled "well enough". That is dangerous belief to those caught-out by

> Consider the impact that simply believing the relevant portions of the world has been modeled "well enough". That is dangerous belief to those caught-out by the edge-cases, especially seeing how actual humans lazily defer to, and others defend "the (application) system": see any HN thread on AI bias and/or Tesla Autopilots deficiencies.