Co-founder&General Partner: The Venture Reality Fund/VRF; investing in Spatial AI & Physical AI 🤖||| @stanford @carnegiemellon @SJSU @NPS_Monterey
The Pleiades
Joined February 2012
- Tweets9.4K
- Following2.5K
- Followers3.7K
- Likes174K
Pinned Tweet
Here’s the book to read for MPC by Profs Bryson “The Father of Modern Control Theory” & Ho
@stanford @StanfordEng
Applied Optimal Control a.co/d/9VsSpmu
Marco DeMiroz ⚽️ retweeted
We're excited to introduce AW-0: Agentic World Action Models! 🤖 (0/8)
The first World Action Model that plans in code, predicts the future, and controls the robot—all in one generative model.
AW-0 jointly generates:
🧠 Code - explicit skill calls, object arguments, and control flow
🎥 Future video - predictive visual dynamics
🦾 Robot actions - continuous motor control
All three are generated together through joint denoising, letting semantic plans directly shape both predicted futures and robot behavior, with:
🤝 Hybrid execution — generated programs can also run directly through perception and control APIs
⚡ Just 40 ms of additional inference latency over the base WAM on an RTX 5090
📚 Pretrained on ~3,000 hours of robot + human demonstrations
One model for planning, prediction, and execution.
🌐 Project: agentic-wam.github.io/
📄 Paper: arxiv.org/abs/2610.30136v1
💻 Code: github.com/ZhuoyangLiu2005/a…
🤗 Model: huggingface.co/miniFranka/ag…
🧵 How it works—and what we found ↓
Marco DeMiroz ⚽️ retweeted
The robotics landscape is fragmented.
We have models, sims, data, evals, tooling, and so much more
@theSamPadilla and I built out the robotics map for public service and industry enthusiasts
If we missed something or got something wrong, let us know at
roboticsmap.org/
Marco DeMiroz ⚽️ retweeted
MASSIVE WEEK in Physical AI.
It’s getting difficult to catch up these days.
Models
@dreamlabs_ai came out of stealth with Dream-1, a robot foundation model available through a public API
@EkaRobotics raised $150M to scale its Vision-Force-Action model for dexterous manipulation
Data
@MaxinsightsAI raised a $30M Seed and launched MaxLoop + MaxGlove
@origami_robo released S1, a skill capture glove for collecting human manipulation data
@PantheonInc open-sourced DiverseUMI-100 with 12,000+ tasks
Simulation
@MidcenturyAI released Akari, a learned rendering model for robotics simulation
Infra
@nebiusai updated its Physical AI Workbench with NVIDIA, connecting data generation, simulation, training and evaluation
@nvidia expanded the Isaac stack with Isaac Lab 3.0, Isaac OS, new datasets, simulation assets and Agile ONE SN
Dreamscale launched Deployments, an inference engine for serving fine-tuned checkpoints on robots in real time
Evals
@xdofai launched the XDOF Robotics Leaderboard for evaluating models on real robots
Industry / Deployment
@Tesla_Optimus started early production of Optimus at Fremont
@agilityrobotics + AGCO signed a commercial agreement to bring Digit into its Kansas factory
Agile Robots + Audi showed mobile humanoids picking and sequencing parts at Audi’s Ingolstadt plant
@NEURARobotics + TUM opened RoboGym, a 2,300+ sqm robot training center, and unveiled 4NE1 Gen 4
Marco DeMiroz ⚽️ retweeted
A researcher at Jeff Bezos's new AI lab put his entire robot learning course on YouTube: 11 lectures from this spring plus 10 guest talks from Physical Intelligence, NVIDIA, Google DeepMind and Meta, free.
"Robot Learning: From Fundamentals to Foundation Models," Oier Mees (@oier_mees), ETH Zürich, Spring 2026.
Robot control and MDPs, imitation learning, reinforcement learning, generative models, transformers, world models, generalist robot policies (VLAs), embodied reasoning - 11 lectures and 10 guest talks, about 17 hours, plus 4 coding homeworks on GitHub.
One guest talk is Quan Vuong, co-founder of Physical Intelligence, on π0.7, the robot brain they released in April. 13 minutes in, it folds a shirt on a robot arm that had no folding data of its own.
Figure is hiring Helix AI Engineers right now, at $200K–$400K. Helix is Figure's VLA, and lecture 9 is how VLAs are built.
Lectures: youtube.com/playlist?list=PL…
Guest talks: youtube.com/playlist?list=PL…
Course: cvg.ethz.ch/lectures/Robot-L…
P.S. the other guests: Cheng Chi (@chichengcc), who built Diffusion Policy, Ted Xiao (@xiao_ted), ex-Google DeepMind, and Lucas Beyer from Meta. my Optimus-style robot brain guide is below if you want the 8-week path.
Marco DeMiroz ⚽️ retweeted
What an amazing resource for anyone interested in robotics!
For years, I've been looking for a textbook that covers the fundamentals of robot autonomy. For my robotics course at @Princeton, I used a mishmash of different textbooks for motion planning, feedback control, SLAM, and robot learning.
The new book Principles of Robot Autonomy from @drmapavone, @DanieleGammelli, Joseph Lorenzetti,
@katielulula, and @GioeleZardini is incredible! It goes all the way from the foundations of planning, control, and SLAM, to modern robot learning (imitation + RL).
I am using this as the primary reference textbook for my Intro to Robotics course at Princeton.
Book: arxiv.org/abs/2608.03496
Lectures: porabook.com/lectures/
Ever wondered what it’s like to learn robot autonomy @Stanford? Now you can.
Today, we’re opening up the robot autonomy teaching materials that my collaborators and I have been developing over the past several years—for anyone, anywhere, completely free.
At the heart of this effort is Principles of Robot Autonomy (PoRA), a textbook inspired by the Stanford robot autonomy course series I’ve been teaching since 2017.
Robot autonomy has evolved from a research discipline into a technology deployed in the real world. Building capable autonomous systems requires more than understanding individual algorithms: it requires understanding how perception, state estimation, planning, control, and learning work together in a complete autonomy stack.
That’s what we set out to teach with PoRA: bringing classical robotics and modern AI together in one coherent framework, grounded in both theory and practice.
Coming soon in print from Cambridge University Press, Principles of Robot Autonomy is also freely available online and comes with a comprehensive collection of open educational resources:
📖 Textbook, arxiv.org/abs/2608.03496 — Foundations of robot autonomy, from classical methods to modern AI-based approaches.
🎥 Video lectures, porabook.com/lectures/ Lectures from my Stanford courses on robot autonomy and on optimal and learning-based control.
💻 Interactive exercises, porabook.com/exercises/ — Hands-on Python implementations and Jupyter notebooks.
📝 Teaching materials, porabook.com/instructors/ — Lecture slides and resources for instructors.
🌐 Explore all PoRA resources, porabook.com/ — Including Ask the Book, an AI study assistant integrated throughout the textbook to answer questions, explain derivations, and connect ideas across chapters.
📣 Read the Stanford announcement: news.stanford.edu/stories/20…
Our goal is simple: to make high-quality education in robot autonomy accessible to students, researchers, engineers, educators, and anyone curious about how autonomous robotic systems work.
I believe that sharing knowledge is fundamental to the mission of a university. Autonomous robotics—and Physical AI more broadly—are among the defining technological frontiers of our time. We want to help more people understand these technologies, build with them, and advance the field.
A huge thank you to my co-authors, @DanieleGammelli, Joseph Lorenzetti, @katielulula, and @GioeleZardini, and to everyone who helped bring this project to life.
Explore the resources, use them in your own learning or teaching, and please share them with anyone who might find them useful!
#Robotics #PhysicalAI #ArtificialIntelligence #AutonomousSystems #Education
@Stanford @StanfordEng @StanfordAILab @StanfordHAI
Marco DeMiroz ⚽️ retweeted
We are releasing our first model, Flint, a best-in-class local OSS coding model that runs at 150+ tokens/s on an @nvidia DGX Spark and 125+ tokens/s on a MacBook Pro M4—developed from scratch in less than four months with help from @Recursive_SI's autoresearch system.
Introducing Flint, @Recursive_SI’s first model: a fast local coding model built with our automated research system, the Eureka Machine.
• Best-in-class fast local coding
• Up to 2x faster than comparable models: 125+ tokens/sec on a MacBook Pro, 150+ on RTX Spark
• Built from scratch in under 4 months
Private, unmetered intelligence on your own machine. Open weights available now on @huggingface, run it locally on your favorite framework or try it for free on @opencode or @pidotdev.
Marco DeMiroz ⚽️ retweeted
Try Flint today.
Read the blog: recursive.com/updates/introd…
Download on Hugging Face: huggingface.co/RecursiveSI/f…
Run Flint locally with @ollama, @lmstudio, @UnslothAI, @exolabs, @Nativ_AI, @sgl_project, @vllm_project and llama.cpp or try it for free on @opencode and @pidotdev
Huge step towards recursive self-improving super intelligence! @Recursive_SI @RichardSocher 👏
Introducing Flint, @Recursive_SI’s first model: a fast local coding model built with our automated research system, the Eureka Machine.
• Best-in-class fast local coding
• Up to 2x faster than comparable models: 125+ tokens/sec on a MacBook Pro, 150+ on RTX Spark
• Built from scratch in under 4 months
Private, unmetered intelligence on your own machine. Open weights available now on @huggingface, run it locally on your favorite framework or try it for free on @opencode or @pidotdev.
Marco DeMiroz ⚽️ retweeted
Introducing Embodied First Exam:
One line of code to run coding agents in 1,090 embodied tasks across 31 simulation suites and 44 robots.
Can your agent go far beyond pick-and-place:
Puzzles with spatial constraints;
Olympic sports with multi-robot collab;
Real-time eval, where the world won't pause;
Dynamic, agile, whole-body control;
AutoResearch with sim2sim
Try your own agent:
harbor run -d embodied-first-exam -a <your_own_harness> -m <your_own_model>
Marco DeMiroz ⚽️ retweeted
Robotics has spent years scaling demonstration data and model size. What happens when we scale EXPLORATION for robot agents?
We're incredibly excited to share RAX, an agentic robotics system that learns by scaling online exploration.
A coding agent connects directly to the robot’s perception and control APIs, learns from feedback, and builds programs that perceive, plan, and act. The model stays frozen. As exploration scales, the agent improves its programs and builds up reusable skills.
The results:
• ZERO teleoperation or action demonstrations on real robots—from insertion with 0.1 mm clearance and air-fryer manipulation to deformable objects and long-horizon mobile whole-body control.
• +37 percentage points on average over leading learned policies across five simulation benchmarks: RoboDojo, BEHAVIOR-1K, HumanoidArena, DexJoCo’s bimanual tasks, and LIBERO-Pro.
• One framework spanning single-arm, bimanual, dexterous, humanoid, and mobile robots, with new embodiments connected through APIs.
The craziest part: on each real-world embodiment, going from scratch to mastering the tasks took roughly two hours. No teleop. No model finetuning.
We’re open-sourcing the full framework AND every experiment session. We can’t wait to see what people build with it!
Project + sessions: robot-agent-exploration.com/
Paper: assets.robot-agent-explorati…
GitHub: N
Project fully led by @RunhanH and @xczhangxx who are absolutely awesome.
How it works, experiments, and robot videos below 🧵
(1/N)
Marco DeMiroz ⚽️ retweeted
list of 15 best robotics researchers on X:
1. @DrJimFan, Jim Fan
2. @chelseabfinn, Chelsea Finn
3. @svlevine, Sergey Levine
4. @pabbeel, Pieter Abbeel
5. @drfeifei, Fei-Fei Li
6. @rodneyabrooks, Rodney Brooks
7. @0x_meden, Martin
8. @yukez, Yuke Zhu
9. @LerrelPinto, Lerrel Pinto
10. @SongShuran, Shuran Song
11. @tonyzzhao, Tony Zhao
12. @YunzhuLiYZ, Yunzhu Li
13. @xiaolonw, Xiaolong Wang
14. @GuanyaShi, Guanya Shi
15. @RemiCadene, Remi Cadene
Marco DeMiroz ⚽️ retweeted
We are open-sourcing V1.0 of our agentic robotics framework, ASPIRE, with full real robot skill discovery and self improvement capabilities! On the day Astra arrived, our lab switched the backend model and saturated all the benchmarks to ~99% over night.
And the frontier is already on the move:
- Opus 5.5 slightly edges out Astra on our robot simulation bench.
- Sonnet 5.5 hits the Pareto frontier of cost vs. robot capability, $12/task.
We first introduced ASPIRE and /robot skill library in June, way before the big models could reliably do robotics. A rising frontier tide lifts all boats, but Astra was a 0->1 tsunami on many tasks.
Now you can also feel the Physical Bitter Lesson at home like we did on that fateful day ;)
links in thread
Marco DeMiroz ⚽️ retweeted
many builders in physical AI are focused on deployment. you may have noticed how little robots are actually deployed by leaders of the space.
we are so, so early.
i've been working on this issue for some time. drop your own experience below, soon to be many in-person group discussions on this topic. here is my current read:
-only 30,000 robots are deployed per year in the US across sectors. mostly programmed, not learned models. china does 10x that. its growing fast, but the obstacles are still there
-the advantages of learned robotic models currently struggle to outperform programmed robotics in industrial environments. expect hybrid solutions to blend the strengths of the two
-contract manufacturing, or high mix would greatly benefit from adaptive robotics, but the contract manufacturing space in the USA is..broken. fixing it to deploy becomes a whole economy solution.
-dependability and capex remain the barriers to entry, building a deployment that survives those plus future regulation, liability, and maintenance is how you win
-most industrial spaces lack the data readiness or network infra to support learned robotics. this is a huge problem that labs have either chosen to solve themselves in a fragmented way or avoid wholesale by adding local compute. big opportunity.
-large labs that have focused on model & hardware development are struggling to divert focus to deployment, for good reasons. if you find a way into market, consider that your moat for now.
-as LLMs, world models compete with classical architectures you will see these labs begin to focus on getting deployment partners. picking the right stack is less important than getting the right customers. treat the bitter lesson as your own litmus for vendor choices.
all the companies you admire or consider to be saturating the opportunity are not even close to scaled deployment, you are sitting on the ground floor. get building.
Marco DeMiroz ⚽️ retweeted
if you sell AI training data, see how you compare: we found 159 companies selling data, RL envs and evals, every fact sourced
plus 38 programs that pay companies for their data: whosellsdata.com
8 of the 14 that pay companies directly publish a dollar range. most of the rest set terms per deal. unknowns are marked and nothing is ranked
the other page: what each program wants, who qualifies and what it offers: whosellsdata.com/buying
missing something, or something wrong? fix it on the site, a person reviews every change
we built this at @Secludy. if you buy, broker or license company data and need the names out before it reaches a lab, dm me
Marco DeMiroz ⚽️ retweeted
Finally, we’re publicly introducing our World Reconstruction Model, built for Physical AI and for any application where hallucinations aren’t acceptable.
The same reconstruction engine covers:
• Real2Sim2Real, evals and benchmarks for robotics
• Free-viewpoint replays in near real time for sports broadcasting
• Real performances in full detail for VFX and film
• A self-service reconstruction suite for 4DGS
Ground-truth detail and sub-millimeter geometric accuracy carry across all of them.
Explore more at gracia.ai
Marco DeMiroz ⚽️ retweeted
Upload robot data, fine-tune our most capable models and connect your robots to cloud hosted inference through the API. We handle the infrastructure so you can focus on what’s unique to your use case.
See what teams are building and join the beta at pi.website/blog/api-beta
Marco DeMiroz ⚽️ retweeted
This is the capital of robotics! 🌁
Silicon Valley is home to so many physical AI companies that you could spend a month visiting them and still not see even 10% of them (trust me, I tried).
So, the Bay Area is the world's leading ecosystem for robotics startups, bringing together top AI talent, top universities, experienced founders, and unmatched access to capital.
The region is anchored by Stanford University and University of California, Berkeley, two of the world's top universities for AI and robotics. They produce a constant stream of researchers, founders, and breakthrough technologies.
The Bay Area is also home to many of the companies shaping the future of robotics and AI, including @Figure_robot, @physical_int, @GeneralistAI, and major AI labs such as @OpenAI, or @GoogleDeepMind.
This concentration of talent makes it easy for startups to recruit experienced engineers and collaborate with leaders in embodied AI.
Not mentioning that it is also home to leaders such as @NVIDIARobotics, whose headquarters and leadership in AI chips power much of today's robotics revolution, and Tesla, whose work on autonomous driving and humanoid robots has created a deep pool of robotics, AI, and manufacturing talent.
Perhaps its biggest advantage is access to capital, talent and ambition.
The Bay Area has the world's deepest network of venture capital firms, serial entrepreneurs, and technical leaders who are willing to fund bold, long-term robotics companies.
In the comments I'll post the companies from the ecosystem.
‼️ Note that Bay Area has >300 robotics companies, research labs, and innovation hubs, so this is a curated selection of the notable product companies, not an exhaustive census!
P.S. I'm constantly working on improving these maps, so if your company is missing, please DM me with basic info about the co, and I will include it in the next release.
P.S.2 I have last open slots for my December RoboTour in the Bay Area, if you want to create content together and show the world what are you cooking, dm! :)
~~
♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com
Marco DeMiroz ⚽️ retweeted
State of Physical AI right now:
1. Capital is pouring in. Startups are coming out of stealth and raising hundreds of millions.
2. We’re moving from demos → deployments. Reliability, uptime and actual ROI are starting to matter more than flashy demos.
3. Specialized robots will likely find PMF much sooner than general-purpose robots. We’re already seeing them deployed across agriculture, factories, mining and industrial inspection with companies like @carbonrobotics, @PathRobotics, @maticrobots and @GeckoRobotics.
4. Physical agents feel like the next frontier. Not just completing a single task, but owning entire physical workflows. @DynaRobotics is doing some really interesting work here.
5. Robot design is getting much more interesting. Form factor should follow the job, not necessarily the human body. Really like dyna's Taku design here.
6. Simulation + evals are becoming a critical part of the stack. Testing, validating and iterating on real hardware is slow and expensive, while measuring real-world reliability is still largely unsolved. Companies like @antiochrobotics are working on this layer.
7. Physical AI is an incredibly difficult industry. Models, hardware, simulation, evals, manufacturing and deployment all need to work together before we see robots operating reliably at scale.
Still early, but feels like we’re entering the most interesting phase of robotics yet.
Marco DeMiroz ⚽️ retweeted
Lecture videos, notes/slides, assignments, and code for the Planning & Control module of Intro to Robotics @Princeton are up:
youtube.com/@intro-to-roboti…
irom-lab.princeton.edu/intro…
Next up: estimation and SLAM. I'll build up from the basics (Bayes filter, particle filter) to VGGT.
