The symbolic logic AI in Agre's writing differs in many technical ways with 2020s deep learning, so it's surprising how on-the-mark the critique still reads. There's still the same deliberate vagueness in how the world gets distilled into symbols and representations to begin with.

Comments