← Multimodal Sources 🕐 2 min read
Multimodal Sources

Training General Robots for Any Task: Physical Intelligence’s Karol Hausman and Tobi Springenberg

> We started deploying them ourselves already.

See also (wiki): wiki/physical-ai-capability-levels.md

Show: Training Data · Publisher: Sequoia Capital · Host: Sonya Huang, Pat Grady

Episode URL: https://pscrb.fm/rss/p/traffic.megaphone.fm/CPUAI2576277674.mp3

Publish date: 2026-01-06
Duration: NAs
Default source credibility: HIGH — Sequoia partners interview frontier-lab founders + F500 AI buyers. VC-hosted — portfolio-company framing on recommendations; named guest metrics stay HIGH. Peer-tier to No Priors in quality.

  • Physical Intelligence’s foundation models for robotics show significant progress in generalizing tasks and improving performance, enabling real-world deployment.
  • The company’s PyStar 0.6 model demonstrates improved reliability and efficiency in tasks like making coffee and folding laundry, running for extended periods without failure.
  • The approach of learning from experience through reinforcement learning is seen as a key step towards continual learning and broader deployment of robotic models.

Extracted quotes

# Credibility Speaker Org Timestamp Topic Quote
1 HIGH Tobi Springenberg (Co-Founder) Physical Intelligence 09:14 01-ai-native-landscape We started deploying them ourselves already. We thought this was something that was going to take something like five years to get to the point where the technology is actually ready to deploy a robot in a commercial setting and have it do something valuable. But we’ve done it, I think, two months ago or something like that.
2 HIGH Tobi Springenberg (Co-Founder) Physical Intelligence 32:40 02-corporate-tools We increased kind of throughput of the policies by over 2x on these three tasks. So there’s one task was this box building task I already talked about. One was the making coffee with an actual kind of industrial scale espresso machine. And the other one was kind of like folding laundry.

Per-quote detail

1. Tobi Springenberg — Physical Intelligence (09:14)

We started deploying them ourselves already. We thought this was something that was going to take something like five years to get to the point where the technology is actually ready to deploy a robot in a commercial setting and have it do something valuable. But we’ve done it, I think, two months ago or something like that.

  • Stat: Physical Intelligence deployed robots commercially within two months, earlier than expected five-year timeline.
  • Credibility: HIGH — Named exec at identifiable org with specific timeline and unscripted interview.
  • Topic tag: 01-ai-native-landscape

2. Tobi Springenberg — Physical Intelligence (32:40)

We increased kind of throughput of the policies by over 2x on these three tasks. So there’s one task was this box building task I already talked about. One was the making coffee with an actual kind of industrial scale espresso machine. And the other one was kind of like folding laundry.

  • Stat: Throughput of policies increased by over 2x on three tasks: box building, making coffee, and folding laundry.
  • Credibility: HIGH — Named exec at identifiable org with specific metric and unscripted interview.
  • Topic tag: 02-corporate-tools

Extracted 2026-04-14T17:33:51 via scripts/podcast_mine.py (MLX mlx-community/Qwen2.5-32B-Instruct-4bit).