Mech Lens

Keeping up with the humanoid robot world.

[Mech Lens @bschreck307]

That's the transfer problem, framed as a product feature. #robot #Humanoid

[Sawyer Merritt @SawyerMerritt]

Figure has introduced their new humanoid robot neural network called Helix 2.5, which the company says is capable of entering homes it has never seen before and immediately performing complex tasks with zero additional training.

Figure says it tested its humanoid across 30 previously unseen homes, where it autonomously:
• Made beds
• Folded towels
• Tidied living rooms
• Navigated unfamiliar layouts
• Manipulated objects it had never encountered
• Self-corrected when it made mistakes

"No data was collected from the homes beforehand, and the robot received no fine-tuning or adaptation for the environments or objects."

[Mech Lens @bschreck307]

It’s amazing how robot hands can now replicate subtle human hand movements. #Dex5S #HumanoidRobot

[Eren Chen @ErenChenAI]

Meet PrimeBOT T1 🤖 Price starts at RMB 19,999 (approx. US$2,960), Overseas pricing TBA.

They are scaling up and producing one humanoid robot off the line every 2.5 minutes, with capacity for 10,000 per month.

[Mech Lens @bschreck307]

This really makes me rethink what AI storytelling can bring to creative industries. #GenerativeAI

[el.cine @EHuanglu]

this AI short tell a better story than most modern movies

[Mech Lens @bschreck307]

I love this testing approach. Instead of a polished lab set, Figure uses regular rental homes to push robot capability.
#HumanoidRobot #Robotics

[Figure @Figure_robot]

Today we’re releasing Helix 2.5

We rented 30 homes in the Bay Area. The robots arrived with no additional training and started doing useful work

[Mech Lens @bschreck307]

This is a big milestone. Getting humanoids to work safely right next to people is what will unlock real-world factory adoption.
#Robotics #Humanoid

[Agility @agilityrobotics]

Introducing Digit 5.

Agility’s next-generation humanoid, engineered for cooperatively safe work at scale, allowing it to work in close proximity to people without the physical safety barriers required by traditional automation.

Explore Digit 5: https://t.co/wurx6x4brm
Watch the full video: https://t.co/fOgRz9bzMd

[Mech Lens @bschreck307]

Cross-embodiment transfer is one of the most important goals for robotics. This result shows the model can adapt well even on different robot hardware.
#AGIBOT #EmbodiedAI

[AGIBOT @AGIBOTofficial]

AGIBOT releases GE-Act 2.0 — the first native World Action Model to validate a pretraining and scaling path for embodied AI.

📖 Explore the project: https://t.co/zMHGNrCbX2

Trained entirely from scratch on embodied manipulation data: visual representation, future generation, and action prediction, all from random initialization. No inherited video generators. No task-specific fine-tuning.

Put straight to a ruthless real-robot zero-shot test — unseen scenes, unseen objects, 100 atomic tasks, 20 skill categories, and two robot embodiments:

✅ Data scaled 100×: from 300 to 30,000 hours
✅ Task success climbs from 17.1% to 44.1% on G1-OP — with no sign of saturation
✅ New skills emerge at scale: folding towels, nesting paper cups, uncapping pens, arranging flowers
✅ Cross-embodiment transfer: G2-90D, under 2% of the training data, still gains 17.7 percentage points
✅ Failure data becomes a training asset — 2,000 hours of failed manipulations and deployment rollouts

A capable model envisions reality before it acts.

#AGIBOT #EmbodiedAI #WorldModel #PhysicalAI #Robotics

[Mech Lens @bschreck307]

Love this creative way to show what these robots are capable of.#AGIBOT #Robot

[AGIBOT @AGIBOTofficial]

🤖 Who said humanoid robots can’t join the circus?

🔥 Jumping through a flaming hoop.
🪀 Pulling off a diabolo trick.
🪢 Jumping rope in sync.
📦 Working together to lift and carry.
What looks like a circus performance is actually a serious test of whether a robot can perceive, react, and move in real time.

Powered by AGILE 2.0, AGIBOT X2 takes on a series of dynamic tasks driven by vision — turning visual perception into real-time, coordinated motion.

Welcome to the AGIBOT Circus. 🎪🔥

#AGIBOT #AGIBOTX2 #HumanoidRobots #EmbodiedAI #Robotics #AI

[Mech Lens @bschreck307]

Curious what kind of real-world tasks these NVIDIA humanoids will be tested on first. #Robotics

[NVIDIA Robotics @NVIDIARobotics]

Building general-purpose humanoid robots requires an end-to-end approach.

Join NVIDIA’s Spencer Huang to see how AI agents and NVIDIA Isaac GR00T are accelerating the entire process.

Add the session to your calendar 👉 https://t.co/QAs0uduPg5

[Mech Lens @bschreck307]

This corn kernel demo really puts the robotic precision into perspective. The potential for surgery is huge. #AIBreakthrough

[Mech Lens @bschreck307]

It’s absolutely mind-bending watching AI agents tackle these legendary unsolved math problems that have stumped human mathematicians for generations. #AIResearch

[OpenAI @OpenAI]

We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics.

The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra.

The problem concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.

[Mech Lens @bschreck307]

This automated production line is more than just a demo. It signals the industry is moving past prototypes and seriously preparing for commercial volumes. #AIrobotics

[Mech Lens @bschreck307]

Watching Astra reliably manipulate physical objects definitely boosts my confidence for general‑purpose robots. We are slowly closing the gap between language models and real‑world physical action.
#EmbodiedAI

[Mech Lens @bschreck307]

Massive GPU partnership from Figure, compute really makes or breaks humanoids.

[Mech Lens @bschreck307]

Badminton is incredibly fast for robotics.
This machine tracks, reacts and strikes the shuttlecock in real‑time.
Great showcase for high‑speed perception & motion control.

[Mech Lens @bschreck307]

While most teams focus on robot bodies, Figure bets big on compute infrastructure.
Nscale + NVIDIA Vera Rubin delivers up to 100k GPUs.
Multi‑billion‑dollar spend to realize home‑use humanoids.

[Figure @Figure_robot]

Today, we’re partnering with Nscale to deploy up to 100,000 GPUs on the NVIDIA Vera Rubin Platform

We're committing $3.5 billion initially, with plans to scale beyond $6 billion

Bringing a robot into every home demands compute at an unprecedented scale

[Mech Lens @bschreck307]

"The biggest product ever" is the marketing frame. how many Optimus units have shipped to non-employees, doing paid work, for weeks on end? That's the real number. Waiting on it.

[Humanoid Intelligence @HumanoidsHQ]

Tesla Optimus will soon be able to perform all of these tasks and much more.

As @elonmusk said:

“Optimus will be the biggest product ever.”

[Mech Lens @bschreck307]

Figure exits stealth with Index: 264K app downloads, 16M videos uploaded. Big claim for scaling general-purpose robots. Data moats are real — whether this one is, depends on what ships next.

[Figure @Figure_robot]

Introducing Index

Today we're coming out of stealth with the largest and most diverse robot dataset to scale general purpose robots

To date we've crossed 264,000 app downloads and uploaded 16 million videos

[Just Mindy 🐊 @just_mindy]

Gas is now $2.59. It was $1.89 the last days of Trump admin. But y’all’s little hearts won’t be offended by mean tweets and that’s what matters. ♥️