No Way Out AI’s D&C Cycles Accelerate: Harmonizing Agents and The Big 'O’rientation with Mahault Albarracin, PhD
7 snips
Sep 4, 2025 
In a thought-provoking discussion, Mahault Albarracin, PhD, Director of Research Strategy at VERSES AI, draws from her social sciences background to explore the intersection of neuroscience and AI. She challenges the notion that more power is the answer, advocating for systems that understand context and think like humans. The conversation dissects the OODA loop and highlights the importance of agency in adaptive systems. Albarracin also critiques prevalent AI doom narratives and underscores the necessity for ethical alignment and meaningful connectivity among intelligent agents.
AI Snips
Chapters
Books
Transcript
Episode notes
From Social Science To Active Inference
- Mahault moved from social sciences into active inference after finding narrative explanations lacked predictive mechanisms.
- She studied ecology, predictive processing, and met advisors who connected her to Pierre Poirier and Carl Friston.
Embedding Beats Scaling Alone
- Current large models focus on attention but miss embodied, spatiotemporal embedding and causal structure.
- Simple structural priors and embodiment yield more meaningful, data-frugal intelligence.
Design Agents With State And Causality
- Build agents with Markov blankets that hold stateful belief spaces and causal models.
- Design agents to navigate and learn causal graphs for predictable, composable behavior.
