Victor Montesinos

Building robots that learn to work—literally— @lutecompany

[Lute @lutecompany]

delightful behaviour comes from the optimized stack - from process handling to tuned kernels

read more on our latest blogpost!

[Lute @lutecompany]

when your infra stack is max out, you unlock restlessness in your robots

more on that soon!

[Lute @lutecompany]

Real deployment is nothing like the demo.
We're hosting a dinner for robotics engineers, integrators & investors — candid conversation, no keynotes.
Warsaw | Lute × Seedcamp × Bones Studio
👉 https://t.co/83SEMnlTth

[Lute @lutecompany]

Your robot is always acting on the past. The only question is how far back. We built an LED clock to measure it: ~17ms of staleness in a live control loop.

[Lute @lutecompany]

Our team at #ICRA2026
We’re already full of inspiration and new ideas and there is so much more to discover.
Let’s discuss robotics together! 🤖

[Lute @lutecompany]

This week we're in Vienna with the brightest minds in robotics, shaping what's next for autonomous systems.

Come find us at #ICRA2026 👋

[mimic @mimicrobotics]

We’re excited to showcase mimic-video, a new class of Video-Action Model, paving the way for scaling robot learning from pure RGB video pretraining. Read more:

[mimic @mimicrobotics]

Today, we’re excited to announce our $16M seed funding round, accelerating our mission to deploy frontier physical AI across industries.

Led by @Elaia_Partners alongside @speedinvest, with participation from Founderful, 1st Kind, 10x Founders, 2100 Ventures, Sequoia Scout Fund.

[mimic @mimicrobotics]

mimic builds on top of years of research in end-to-end models for dexterous manipulation. Today, we’re proud to publish mimic-one, outlining the foundations of our autonomous stack. Check out the paper:

[Ethan Perez @EthanJPerez]

🤖🧘 We trained a humanoid robot to do yoga based on simple natural language prompts like "a humanoid robot kneeling" or "a humanoid robot doing splits." How? We use a Vision-Language Model (VLM) as a reward model. Larger VLM = better reward model. https://t.co/4ujAFCliQa 👇