Loading activity…
@ken_goldberg
Artist and Engineering Prof @UCBerkeley, Co-Founder @AmbiRobotics & @JacobiRobotics. Interests: robots, rockets, redwoods, rebels.
AI
A post highlights a late-breaking results session for work on robotics agents or Real2Sim2Real.
AI
The announcement claims Claude and GPT each built a full-dynamics digital twin from a phone video in an hour.
AI
A post asks for contacts at small or medium manufacturing companies using the approach to improve real production.
AI
Its developer says wrist cameras are critical for robot performance but add bulk, cost and complexity.
AI
Thrun thanked Encord for telling his friends in San Francisco about his appearance.
AI
A September 7 post previewed the 1:30–2:15 p.m. main-stage plenary. A listed participant later invited attendees to Hall C.
AI
Goldberg invited PhD students nearing graduation to discuss collaborating with him and his students on agentic robotics for bimanual manipulation at UC Berkeley’s AUTOLab.
AI
Ken Goldberg says his team plans to work with Nvidia’s team to integrate Robot Sim Studio (RSS) with Newton for direct use by frontier LLMs.
AI
UC Berkeley roboticist Ken Goldberg says his team used Astra to build a physics simulation from a short video of a robot wiping ink with a sponge, then developed control code that worked on a real robot.
AI
A September 28 post advertised tours at 2 p.m. Monday through Wednesday that week, with free tickets to reserve through Eventbrite.
AI
A user recalls reconnecting over a conversation about the historical progression from Jester to Hot-or-Not, Facebook and YouTube.
AI
A post previewing IROS 2026 says two UC Berkeley undergraduates will present the active-tracing algorithm on September 29.
AI
A post describes AI agents designing and fine-tuning robot control systems, with the aim of improving how e-commerce packages are sorted.
AI
A user sees large-scale real2sim—bringing the real world into simulation—as within reach, noting how hard it once was to imagine OpenAI putting serious resources into 3D/4D.
AI
One post agrees with the view that “training a specialized model in a data-driven way may already have been overrun.”
AI
The post says the authors have no previous publications and asks, “Why is it anonymous?”
AI
A reply says language models could improve with a closed feedback loop, accurate task checks and safety guardrails—but argues the system needed for that setup is still far off.
AI
The plenary debate is scheduled for September 30, 2026, from 1:30–2:15 p.m. on the main stage, according to the post.
AI
@scaling01, operator of LisanBench benchmark, made the statement about future AI models.
AI
Posts question whether a robot video shows teleoperation or true end-to-end autonomy.
