Event
The Physical Intelligence Layer — deployment results reported with Weave and Ultra using π 0.6
Key points
- Weave reports live laundromat deployments in San Francisco and quantifies improvements when π 0.6 is used (vs π 0.5), including fewer missed grasp sequences and fewer interventions when Weave pre-training data is.
- Ultra reports live customer deployments packaging real orders, including a full-shift run at 96.4% autonomy and throughput improvements when Ultra pre-training data is included.
- Post frames a reusable "physical intelligence layer" where deployed robotics teams interface their hardware to foundation models and iterate via real-world data loops.
Company context
Builds warehouse robots designed to slot into existing workstations and automate e-commerce order packaging. Reported as deployed in live customer warehouses packaging real orders; uses human-in-the-loop intervention to maintain correctness while generating training data.
Context
- Company
- Ultra
- Segment
- Logistics
- Event type
- Market Signal
- Geography
- Brooklyn, New York · United States