Latest / Logosofia no Dia a Dia
Pragmatic Insincerity: Why AI Still Doesn’t Get the Joke
Can a machine truly understand why a joke is funny, or is it just calculating the probability of a punchline? In this episode, we dive into the "sarcasm gap" and the new multi-agent frameworks designed to help AI navigate the complex world of human humor and idioms. We examine the technical hurdles of teaching machines to parse "pragmatic insincerity," from the visual wit of New Yorker cartoons to the high-stakes risks of misinterpreting diplomatic cables. Discover why the current "C-minus" performance of frontier models matters for everything from automated hiring filters to national…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.