@laneburgett: While @SpaceX was launching rockets we were using @Starlink to remotely inference our excavator robot model that we tra…
Summary
Lane Burgett shares how they used Starlink to remotely run an excavator robot model trained on 2.5 hours of operator data, based on π0.5 from Physical Intelligence, teaching heavy machines real-world tasks.
View Cached Full Text
Cached at: 05/25/26, 08:43 PM
While @SpaceX was launching rockets we were using @Starlink to remotely inference our excavator robot model that we trained with 2.5 hours of operator data.
We are teaching heavy machines to do real tasks on job sites by learning from real operators in the wild. Soon, we will be expanding to more tasks and machines. Our model is built off of @physical_int’s π0.5.
Read more here: https://labs.actor/research/excavator-in-the-wild…
Similar Articles
@SpaceX: Track Starship in space →
SpaceX announced a way to track Starship in space via a link.
@SpaceX: SpaceX was founded to make life multiplanetary. We’ve been able to expand that mission with our Starlink constellation …
SpaceX promotes its mission to make life multiplanetary, highlighting its Starlink constellation and an AI solution, with a link to learn more.
@elonmusk: SpaceX AI Satellites
Elon Musk announces SpaceX AI satellites, likely referring to Starlink or new satellite capabilities with integrated AI.
@elonmusk: Starlink
Elon Musk mentions Starlink, SpaceX's satellite internet service.
SpaceX has almost finished writing V1.0 of an in-house AI training stack in C (2 minute read)
SpaceX is finalizing a custom AI training stack written in C, utilizing pipeline parallelism and 220k GB300 GPUs to achieve over an order of magnitude speed improvement, with plans to develop an inference stack for reinforcement learning.