Stefan Welker

Google Deepmind Robotics

[Stefan Welker @tek2222]

@hegemon1984 @MartinLukasMD @adcock_brett this video is CGI. Its done with Movella Motion suits and Blender and Unreal engine, theres a video on youtube.

[Pete Florence @peteflorence]

One way of thinking about these results — this is the widest diversity of complex tasks I’ve *ever* seen performed by *any* robot.

Finally, something actually exceeds the ~2010 PR1 videos :)

Also to clarify the video below is teleop, but they have autonomous results for a smaller set but still impressive mix of tasks.

Amazing work @zipengfu @tonyzzhao @chelseabfinn

[Stefan Welker @tek2222]

@mrdoob better support for newer hardware, most companies write their drivers for windows, and put a tiny team on supporting linux half-heartedly.

[Quan Vuong @QuanVng]

RT-X: generalist AI models lead to 50% improvement over RT-1 and 3x improvement over RT-2, our previous best models. 🔥🥳🧵

Project website: https://t.co/GAlvFdqwx5

[Stefan Welker @tek2222]

@phoboslab Chrome Gamepad support is pretty good !

[Stefan Welker @tek2222]

@phoboslab love that game's input mechanics and soundtrack.

[Shuran Song @SongShuran]

@haqhuy’s new project:
*Scaling up* robot data collection using LLM for ✅ task decomposition ✅ reward formulation
*Distill down* into visuomotor policies that ✅ operate from raw sensory input ✅ improve overtime.

Check out the engaging Q&A here 😉 https://t.co/fDkTSi7843

[Tuomas Haarnoja @haarnoja]

We just released our work on robot soccer. I've been working on this for quite some time with my amazing colleagues at DeepMind. It's exciting how deep RL can produce such beautiful behaviors with low-cost robots. Full paper is available at https://t.co/A65rPLL7UD Enjoy!

[Stefan Welker @tek2222]

@haarnoja amazing work we could never do that at robocup.

[Stefan Welker @tek2222]

@tehlike my kids use minecraft for heating.

[Stefan Welker @tek2222]

@karasusan Hi i wanted to contact you

[Andy Zeng @andyzengineer]

We built PaLM-E 🌴🤖 one of the largest multimodal language models to date, trained end-to-end on robot data. Images, text, state inputs, neural scene embeddings – you name it.

And it's fantastic on robots. Check out Danny's thread 👇

[Stefan Welker @tek2222]

@sdamico should all fit into an atmel , no ?

[Stefan Welker @tek2222]

@jmdenisme @Blender wow im going to print this.

[Stefan Welker @tek2222]

@tehlike I always liked clippy, i think the printing animation was pretty genius.

[Stefan Welker @tek2222]

@ID_AA_Carmack or just fix them with 4 balls of PU foam from a spray can

[Stefan Welker @tek2222]

@tehlike free messaging vs paid sms ?

[Stefan Welker @tek2222]

@natfriedman sound effect generation

[Karol Hausman @hausman_k]

SayCan won the Special Innovation Award #CoRL2022

[Stefan Welker @tek2222]

@tehlike If its really in Wyoming, then you could do with peanuts !