Mariano Phielipp

Driving the Development of Visual Language Action Models for Next-Generation Humanoid Robots. Views are my own.

[Barack Obama @BarackObama]

The victory of the opposition in Hungary yesterday, like the Polish election in 2023, is a victory for democracy, not just in Europe but around the world. Most of all, it’s a testament to the resilience and determination of the Hungarian people – and a reminder to all of us to keep striving for fairness, equality and the rule of law.

[Massimiliano Viola @massiviola01]

Thread on VJEPA 2.1🤟

This DEFINITELY flew under the radar: just a few days ago, @AIatMeta released V-JEPA 2.1, taking a massive step toward closing the gap between image and video domains.

For a long time, image backbones were the only option for solving dense vision tasks. This model disagrees, showing that universal spatial understanding also emerges from large-scale video models!🎥

[Mariano Phielipp @mphielipp]

@Hongyu_Lii Interesting work. How you see the work of Florence, Manuelli, Redrake in the context of this work Dense Object Nets: Learning Dense Visual Object Descriptors By and For Robotic Manipulation Florence,Manuelli, Russ Tedrake?

[Mariano Phielipp @mphielipp]

https://t.co/wUqzB4WD1w

[Mariano Phielipp @mphielipp]

@hellorobotinc The work at Meka Robotics was years ahead of its time. Those designs were some of the coolest in humanoid robotics — and given the current state of humanoid robotics, it would be great to see some of those awesome designs make a comeback someday.

[Mariano Phielipp @mphielipp]

"We do this not because it is easy, but because we thought it would be easy" @sundayrobotics 😂

[Xuanbin Peng @Xuanbin_Peng]

What if a humanoid robot could choose how to interact with the environment 🤖 —
soft when it needs compliance, stiff when it needs precision, and force-aware when it must push/pull?

That’s exactly what our Heterogeneous Meta-Control (HMC) framework enables.

Our new framework learns to blend and route across position, impedance, and hybrid control from real demonstrations, letting a humanoid wipe, push, lift, and insert with smoothness and stability in contact-rich & forceful scenarios.

Website: 👉 https://t.co/4KwgN9ehw3

[Sunday @sundayrobotics]

After 18 months in stealth, dozens of prototypes, millions of real-home demonstrations, and one final all-nighter, we’re thrilled for you to say hello to Memo

[Generalist @GeneralistAI]

Read more in our blog post, including early notes from large-scale ablations on our pretraining data.

Blog: https://t.co/qTMkptf3p8

[Mariano Phielipp @mphielipp]

@danijarh @wilson1yan Congratulations @danijarh ! Wonderful work! I’ve been looking forward to Dreamer 4 arriving :) I look forward to seeing it increasingly applied in robotics and the results that follow.

[Mariano Phielipp @mphielipp]

Congratulations to Christopher Agia, Rohan Sinha, Marco Pavone, and the entire CUPiD team for winning Best Paper at the RSS 2025 RoboEval Workshop! 🎉 (https://t.co/ObsaYA26PA) .

[Rohan Sinha @RohanSinhaSU]

Excited to share CUPID 💘 won the Best Paper Award at the #RSS2025 RoboEval workshop!

TL;DR: In 🤖 imitation learning, data quality is crucial. CUPID 💘 directly identifies “good” robot data, i.e., whether a demo will improve success rates. ➡️ big gains via data curation! 🧵👇

[Mariano Phielipp @mphielipp]

🚀 Just released an implementation of CUPiD (https://t.co/EeVDT00Si6)—by the excellent work by Christopher Agia et al. Official code isn’t out yet, so I decided to give it a try: https://t.co/4MZHyC1AV3 🧪

[Mariano Phielipp @mphielipp]

@sarahookr Are you going to or speaking at any interesting events?

[Nando de Freitas @NandoDF]

RL is not all you need, nor attention nor Bayesianism nor free energy minimisation, nor an age of first person experience. Such statements are propaganda.

You need thousands of people working hard on data pipelines, scaling infrastructure, HPC, apps with feedback to drive benchmarks and data, tons of research and engineering on generative models, data mixtures, ablations, RL/selftraining, etc etc and we will probably need lots of people working hard to figure out safety, causal world models, awareness, models that create abstractions comparable to infinity and zero and use these to predict the existence of things like black holes and suggest experiments to verify such hypothesis, or come up with novel engineering designs to generate energy more efficiently, robotics, etc etc.

It takes thousands of people and many ideas. In the end some simple ideas might become obvious but such obviousness only happens in retrospect. Yes, there is a bitter lesson but if we had followed it, we’d still be doing linear regression with RL. Let’s not oversimplify, but rather honour the research and engineering of thousands of people.

Also, people keep rewriting history. When our language understanding start up (darkbluelabs) was acquired by Google about 10 years ago, we joined DeepMind, where the AGI documents were all about concepts, RL, episodic memories and made it clear that there was no room for language. To be honest, back then such a position wasn’t so crazy. Now it seems silly, but only because of the benefit of hindsight.

There’s no 1 or 10 heroes in the history of AI. There’s many 1000s of hard working students, profs, engineers, operations and support people, product folks, managers, even hedge funds among others. Let’s honour the whole community and not just ceos or the philosophers of Bayes, RL, deep learning, etc. I look forward to learning from the next generation and seeing what they will achieve. To them: Don’t buy the existing narratives blindly, innovate. Remember that just like mathematics, AI will advance one grave at the time.

[Mariano Phielipp @mphielipp]

Please robot… I’m out of toilet paper.
No yelling. No awkward moments.
Just a smooth, silent rescue. 🧻🤖😂
#robotics #AI #robotsdoingthings #vla #funrobotics

[Mariano Phielipp @mphielipp]

🚀 Exciting Opportunity! 🚀

https://t.co/yUuncMPVTm

[Mariano Phielipp @mphielipp]

import brain
brain.loading("executive_function")

# DEBUG: Insufficient sleep detected. Retrying...
# DEBUG: Compensatory mechanisms activated (Efficiency -20%)

https://t.co/A8mNWv7cTH("today_tasks") ....

[Mariano Phielipp @mphielipp]

@allenzren @physical_int Nice! Thinking in adding https://t.co/5pHykPPqGB?

[Mariano Phielipp @mphielipp]

nGPT: Normalized Transformer with Representation Learning on the Hypersphere. https://t.co/6cuojTKXFG. Remarkable efficient. (reducing the number of training steps required to achieve the same accuracy by a factor of 4 to 20)

[Mariano Phielipp @mphielipp]

Popular information https://t.co/KHk5rRrbze