the challenge is not collecting more data, it is rewarding the right data asked Chris from @axisrobotics how onchain actually helps with physical AI data. two things: provenance and incentives. contributors globally do eligible tasks, submit trajectories, then go through verification and quality scoring. better quality gets a higher share of rewards. without a shared trust layer you do not know where the data came from, who generated it, under which task, or how it was verified. everything is recorded onchain. that is where crypto is actually useful for physical AI.