Why Robots Need More Than Data to Learn in the Real World: Çetin Meriçli

Field notes

Çetin Meriçli, Co-founder of Atlas Robotics, reveals why the biggest breakthrough in embodied intelligence is not raw speed or flashy demos, but robots that can improve from feedback the way humans do.

If you care about robotics, AI, or where physical work is headed next, this conversation will change how you think about what machines should become.

Why “day one competence” matters more than trying to train robots from scratch
How corrective feedback can teach robots faster than endless trial-and-error
Why the old split between symbolic AI and pure data-driven learning is limiting robotics progress
How continuous learning could make everyday robots, like a vacuum cleaner, progressively better in your own home
Why natural language plus large models may finally make apprenticeship-style robot learning practical