Why humanoid robot demos still fail the generalization test

Jeanine Sinanan-Singh, director of GenAI analysis at Appen.

In Episode 265 of The Robotic Report Podcast, our visitor is Jeanine Sinanan-Singh, director of generative AI Analysis at Appen. She discusses the persevering with significance of people within the loop for coaching robots.

Sinanan-Singh works on agentic evals, RL environments, and dataset design for post-training frontier AI fashions. She additionally explains why spectacular demonstrations don’t essentially show that robotic habits may be generalized.

Sinanan-Singh beforehand based a pharmacy automation startup constructed on a 3D-printing platform to fabricate customized drugs. A Harvard graduate with prior expertise at Microsoft and Surge, she is a frequent voice on AI analysis and reasoning.


Present timeline


Information of the week


Sponsors

Closing the loop of autonomy takes greater than an excellent robotic.

WiBotic offers dependable wi-fi charging and fleet vitality administration for autonomous robots and different battery-powered programs.

This patented know-how offers robots extra freedom to cost all through their regular workflow, serving to enhance the uptime, ROI, and security of your fleet.

AMRs, drones, ground scrubbers, humanoids, and all the pieces in between, WiBotic makes charging easy, seamless, and scalable.

Be taught extra at


To be taught how one can sponsor a future episode of The Robotic Report Podcast, contact Jami Brownlee at jbrownlee@arrowfly.com

Would you prefer to be a visitor on a future episode? Contact Mike Oitzman at moitzman@arrowfly.com



SITE AD for the 2026 RoboBusiness call for speakers
Register now and assist us have fun 20 years of RoboBusiness!

The publish Why humanoid robotic demos nonetheless fail the generalization check appeared first on The Robotic Report.