Ezra was right that some of the things the guest says in this episode, or the way he says them, are irksome. However, I found myself pausing and rewinding several times to ensure I wasn't conflating the messenger with the message, or reacting to surprising claims with knee-jerk reactions, or interpreting words too literally, or even mixing up aspirations with practical strategy. I found this exercise to be illuminating. Overall, it was a refreshing experience to listen to a conversation between two thoughtful and articulate people espousing two *actually* different sides of a complex issue, whereas in many other similar conversations the interlocutor from the "other" side is way too abstract or distant from the core claims being (literally and figuratively) fought over on the ground. Real food for thought!

Not the most insightful discussion on AI I've listened to recently, but it was an interesting listen, and did offer a couple of novel perspectives. I particularly liked the comparison of the engineer/programmer mindset as opposed to the librarian/researcher one for the people most able to thrive in working with the current generation of LLMs (from around 14:40), as well as the manager/editor analogy providing a mental model and frame of reference of how we're most likely going to have to learn to interact with this technology (the ending remarks at around 55:00).

Este é provavelmente o episódio mais marcante que já ouvi de um podcast em português. É daquelas conversas que restauram a fé na humanidade e inspiram a procura de uma vida com significado. Muito bom!