← Home
TWIML · December 17, 2025 · 55m
Building True Agentic AI with Reflection
Aakanksha Chowdhery from Reflection explores the fundamental shifts required to build true agentic AI that can plan, act, and learn autonomously.
This summary was generated from show notes and public descriptions, not from a full transcript review. Details may contain inaccuracies.
Canon
•
Chowdhery argues that current AI agents fail because their environment interfaces are too limited. True agents need rich perception (seeing the environment) and rich action spaces (changing the environment).
•
Chowdhery applies Stoic reasoning to her research timeline: she cannot control when AGI-level agents will emerge. She can control the rigor of her research methodology and the safety of her evaluation frameworks.
Was this useful?