C.K. Wolfe ✈️ CoRL

editor-in-chief of @berkeley_ai blog | cs phd @berkeley_ai ai robotics | former team lead @airacingtech |

[Tara Sadjadpour @TaraSadjadpour]

One human demonstration. Any multi-fingered hand. Zero-shot sim-to-real visuomotor policy.

https://t.co/fqkm4Fc1qv

Collaborators: @he_siming @ckwolfeofficial @HaozhiQ @LeaMue27 Shankar Sastry, Claire Tomlin, @JitendraMalikCV

[Rosario Scalise @rosario_scaling]

Our work is up on alphaxiv! (finally uploaded properly without artifacts) https://t.co/v4CVc26CMG

@octi_zhang @patrickhyin @abhishekunique7 @mateoguaman @iggydagnino !

[Tensr @tensrinc]

From factory floors to orbit.

Excited to be working with @ISS_CASIS as we prepare to unveil Tensr’s orbital robotics projects in 2027.

More to come.

[Rosario Scalise @rosario_scaling]

Huge fan of the work @TaraSadjadpour has been doing all summer! For the newcomers to Physical AI: targeting 3 distinctly different hardware morphologies (even in sim) and getting full sim2real on the fullstack arm+hand deployment for @SharpaRobotics hand...that takes dedication.

Excited to see more of this from the controls/learning dream team at @UCBerkeley EMBER, @he_siming, @ckwolfeofficial, @JitendraMalikCV, Shankar, Claire &co!

@HaozhiQ & @LeaMue27, keep rockin' it at FAR/@UChicago, and @AIatMeta respectively!

[Siming He @he_siming]

🎉 Workshop on Sim-to-Real-to-Field: The Science of Transfer is coming to #CoRL2026 in Austin! Our CFP is now open — with a $1,000 Best Paper Award generously sponsored by @tensrinc .

🔗 https://t.co/W8O36OQiez

[Jitendra MALIK @JitendraMalikCV]

Three data sources are popular for robot learning today (1) Humans tele-operating robots (2) Humans acting as humans doing tasks (3) Robots acting in the "safe space" of a simulator. Happy to share this work from UC Berkeley which shows the power of (2) combined with (3). Visual imitation followed by trial and error is a good recipe for robots as it is for human children. Whither tele-op?

[C.K. Wolfe ✈️ CoRL @ckwolfeofficial]

@TaraSadjadpour @he_siming @HaozhiQ @LeaMue27 @JitendraMalikCV really proud to have contributed to this one, huge work from @TaraSadjadpour + team!

genuinely worth the read.

[C.K. Wolfe ✈️ CoRL @ckwolfeofficial]

@JayHe748646 @chris_j_paxton @nimaarahmanian

[Thomas Wolf @Thom_Wolf]

We have a huge news to share today!

Today we are unveiling the first truly accessible RL robot - welcome Microduck

A 25 cm tiny open-source biped with 15 actuators and packed with sensors (camera, speaker, LiDAR, NFC, bluetooth, wifi, etc) that you train yourself with reinforcement learning.

It's also playable out of the box with more than half a dozen fun and playful pre-trained policies to have it walk, sit, crouch, roller-skate, pick up objects with its articulated beak, and recover on its own.

And all for less than $400.

See all the details, play with the simulator and order it at: https://t.co/n1Btgs6vKw

(video with sound on 🔊)

[Skild AI @SkildAI]

Introducing S1, our new foundation model that learns from one example.

It can be taught 10-minute long tasks that it has never seen before, from one video prompt without any fine-tuning.

Watch S1 operate in real-time via in-context learning:

[Sergey Levine @svlevine]

We can think of this as the robot analogue of RL for thinking, optimizing for good "thoughts" through trial-and-error. The surprising thing is that it's so fast, learning in under a hundred real-world trials.

Website: https://t.co/APpoOIsXB2
Paper: https://t.co/50dkjqN8y3

[C.K. Wolfe ✈️ CoRL @ckwolfeofficial]

Released on @berkeley_ai blog, new work from Jakob Bjorner, @a_lidayan, @satvikgolechha, @kartik_goyal_ & @alsuhr. As task horizons grow, LLM contexts can’t scale forever — ABBEL trains agents to maintain graded natural-language belief states instead of drowning in history … 🐻📄

https://t.co/Vau1yLEjq0

[Berkeley AI Research @berkeley_ai]

We could not be prouder of this years BAIR PhD graduates! Follow their journey to see where they go next @baifeng_shi @sea_snell @databoydg @enfleisig @graceluo_ @zhuhl98 @HaozhiQ @jdzamfi @LianJiachen @joshminwookang @junhaobearxiong @kaylolittlejohn @KentKChang @kvablack @KunheYang @lisabdunlap @LongTonyLian @maulik__bhatt @michaelpsenka @nathanlichtle @neerjathakkar @nikita_mehandru @NiklasLauffer @qiyang_li @vbenara05 @vonekels @xiuyu_l @Yiheng_Li_Cal @ZheFu1 Sampada Deglurkar Wei-Jer Chang Yichen Xie Yigit Erginbas

https://t.co/6axNQ64vDg

[Wenli Xiao @_wenlixiao]

Autoresearch just left the sandbox and entered the embodied world.

We are excited to introduce 𝐄𝐍𝐏𝐈𝐑𝐄: a system that drops frontier coding agents onto a fleet of real robots and hands them the entire loop:

reset the environment → search the literature → implement ideas and build the infra → train and deploy → self-verify → analyze the logs and rewrite the code → repeat, until the policy is reliable in the real world. No human in the loop.

Guided only by the robot's self-proposed, heuristic-based success signal, the agents hill-climb to 99% on dexterous real-world tasks: organizing pins into a box, seating GPUs, tying zip-ties.

We envision the bottleneck in robotics shifting — from building smarter algorithms to building the closed physical feedback loops an agent can finally turn on its own.

🔗 https://t.co/3tL2ArGo3v
From @NVIDIA @CMU_Robotics @Berkeley_AI
🧵

[Matei Zaharia @matei_zaharia]

Really excited to open source a new project: Omnigent, a meta-harness for AI agents.

It lets you build multi-agent coding and custom agents, sitting above Claude Code, Codex, Pi, and agent SDKs to let you compose them. It also adds live collaboration and rich control policies.

[Carlo Sferrazza @carlo_sferrazza]

Good news at ICRA today! Really honored, and grateful to work with a fantastic team 🙂

More about the paper here: https://t.co/gDjuO1eMci

[Rutav @rutavms]

Can we build generalist robots with zero teleoperation? Come participate in the discussion and weigh in at our ICRA'26 workshop, BeyondTeleop, starting at 8.45 am CEST today (June 5th)!

📍 Strauss 3

[Stephen Xie @stephenx_]

Longer chain-of-thought = slower inference, more context rot, and ballooning compute.

So what if the model could decide for itself when to go parallel?

Our new BAIR blog breaks down Adaptive Parallel Reasoning (APR) — the next paradigm in inference-time scaling. 🧵

[CSGE @berkeley_csge]

Berkeley CS Graduate Entrepreneurs (CSGE) is back with the annual Spring Mixer on May 8th! 🌉

Join us for a night where research meets startups, featuring an exciting panel with @sarahookr, @ericzelikman, and @NaveenGRao!

RSVP early to save your spot:
https://t.co/mCIuYuQozv

[David McAllister @davidrmcall]

We developed a simple, sample-efficient online RL technique for post-training image generation models. We see it as a possible steerable alternative to CFG, driven by any scalar reward, including human preference.