The two paradigms In this channel's material, the symbolic vs sub-symbolic AI distinction names a major historical shift in how AI systems are built. Symbolic AI encodes knowledge as explicit, human-readable rules and symbols that people write by hand. Sub-symbolic AI — the approach behind today's neural networks and large models — does not encode knowledge that way; instead, capability emerges from patterns captured in vast amounts of data, and the internal workings are not directly readable as rules. How the channel uses it The term appears in the channel's coverage of Reid Hoffman's Superagency as one of the side threads of preparing for AGI: the shift from symbolic to sub-symbolic AI is part of the transition the material asks readers to understand and prepare for. Hoffman's broader framing — learn from past technology panics, treat agency as partly an internal stance, plan with a compass instead of a blueprint — is the context in which this paradigm shift is raised. - Symbolic AI: knowledge written as explicit rules and symbols by humans. - Sub-symbolic AI: capability learned from data, not spelled out in rules — the paradigm behind modern models. - The channel presents the shift from symbolic to sub-symbolic AI as a thread in Reid Hoffman's argument for engaging seriously with AI now.