Dyna-2 uses over a million hours of human video to cut how much robot data each task needs, and its robots are already working in restaurants.
October 14, 2026 edition: Robotics sector, 1 storyRobotics: latest AI analysis
Robotics on AT Pulse covers machines in the physical world. The desk has published 8 stories tagged Robotics across 7 daily editions, from September 16, 2026 to October 14, 2026, citing 94 linked references. The most frequent original sources are TechCrunch AI, Latent Space, and Machine Learning Street Talk · podcast. Stories here most often overlap with Models, Business, and Research.
- stories
- 8
- editions
- 7
- references
- 94
- latest
First story September 16, 2026. Counts come from the published archive and update with each edition. Summaries are AI-assisted; see our methodology.
Last updated
RECENT EDITION VIEWS
What recent editions said about Robotics.
World models and cheaper humanoids are moving toward real deployment and entertainment. That raises questions about adapting models across machines and about how people respond to robots.
October 12, 2026 edition: Robotics sector, 2 storiesRobot reliability and scarce training data still limit home humanoids, and forecasts for when they arrive vary widely.
October 8, 2026 edition: Robotics sector, 1 storyLATEST
8 of 8Newest Robotics stories.
Podcast · Robotics
Dyna Robotics unveils Dyna-2, a robot world-action model pretrained on over a million hours of human video
Dyna Robotics' Dyna-2 predicts future video and robot actions together. It was pretrained on more than a million hours of first-person human video.
Models
Odyssey opens free browser preview of Odyssey-3, a world model pitched for robots, cars and agent training
Odyssey-3 is a single world model that Odyssey says can be adapted to robot arms, humanoids, cars and agent training environments with tens of hours of data.
Safety
TechCrunch essay examines why people treat AI as human, drawing on Sherry Turkle's new book 'Artificial Intimacy'
A TechCrunch essay starts with a reporter shaking a Unitree humanoid's hand at MIT CSAIL and moves to Sherry Turkle's argument that chatbots should keep to bounded tasks.
Policy
Anthropic's new usage policy bans needless cruelty toward Claude and lifts its blanket ban on campaign targeting
Anthropic's revised usage policy, effective November 12, 2026, bans sustained, purposeless abuse of Claude and enforces it mainly by ending chats.
Robotics
Robot AI's 'unseen task' claims outpace the evidence as forecasts for home humanoids diverge
Physical Intelligence said π0.7 attempted a task it had never seen, but related examples turned up in its training data.
Podcast · Robotics
Gemini Robotics 2 controls whole humanoid bodies, but DeepMind's research lead says moving reliably to new robots is still unsolved
Gemini Robotics 2 pairs a publicly available reasoning model, ER 2, with partner-only action models that control entire humanoid bodies.
Models
World model startups draw major funding while keeping product plans and evaluations opaque
World model companies are attracting large funding while revealing little about products or timelines.
Podcast · Models
NVIDIA’s Cosmos 3 positions omnimodal world models as policy-verification infrastructure for physical AI
Ming-Yu Liu’s interview frames Cosmos 3 as a model family linking language, video, audio and action, with near-term value in synthetic data, post-training and policy checkpoint triage—while leaving real-world validation and safety assurance unresolved.
Stay curious.
Follow the evidence.
Independent perspectives, the original sources, and room for the questions that don't have easy answers.
Connect with us
Find us where you already read.
- X
Short threads on the day’s biggest shifts, with the sources attached.
Follow @atpulsehq - LinkedIn newsletter
The weekly roundup, read it where you work.
Read the newsletter - Instagram
Quick visual takes on the day’s biggest AI shifts.
Follow on Instagram - Email
Tips, corrections and ideas. Notes go to the operator of the AT Pulse pipeline.
at.pulse.ai@gmail.comEmail us