On January 19th, Lightwheel Intelligence, which focuses on simulation technology, announced a deep cooperation with World Labs founded by AI guru Li Feifei to jointly solve the long-standing problem of large-scale evaluation in the industry.
With the improvement of robot basic model capabilities, the bottleneck has shifted from training to evaluation: to measure the true progress of the model, it needs to be tested in thousands of environments, tasks, and failure modes. However, real-world evaluation is slow, costly, and difficult to replicate, making simulation the only feasible path.
Featured article series
This content is part of a featured article series, sign in to access the full article
