A lot of engineering goes into making autonomy look simple Here’s a 2 hour timelapse of F.03 repeatedly walking up and down stairs. Figure HQ is full of tests like this, each one helping push robots closer to fully autonomous systems

@Figure_robot Why does he take the stairs when there is a short-cut? /s

@Figure_robot Not smart enough yet

What’s my purpose? Do 10 thousand reps of stairs

"What is my purpose?", "To walk up and down stairs", "Oh God". 😅 But seriously, this a major part of how robots are going to learn how to do things: learning by doing. There are no shortcuts.

Given how prevalent steps are, maybe it's worth building a rig that can easily adjust step height and width. I'm guessing ~99.99% of steps are between 3"-12" tall and between 7"-24" wide.

Brett, can the stairs introduce variation that would be helpful for 03 to learn? Ie changing the flooring texture or angles faintly etc?

This is wild. 2 hours of stair practice on repeat - the progress on these robots is crazy.

okay😒

I love these kind of posts, keep it up!

Now make 2 of them carry a sofa up and down the stairs

Make the stair heights uneven, then give the robot two bags of groceries to carry.

that's the problem right there - you're treating it like a video game and only doing the observer draw - you have to do the environmental retention in ADDITION to the observer draw

What’s the ETA?

The interesting part isn't the walking, it's the repetition count. Every cycle generates training data that compounds. The humanoid robotics thesis lives or dies on unit economics: whether companies like Figure can get cost per unit down enough to beat human labor in logistics.

Breakthroughs often look sudden from the outside. From the inside, they're usually thousands of repetitions, small improvements, and relentless iteration. Progress is often just persistence made invisible.

Why not use synthetic data? Not good enough? Can't simulate edge cases?

Blazing!

Amazing.

Why does it need to walk so far away from the stairs? And why not try to make it change direction in the middle or make other more difficult moves for example?

@Figure_robot whoa! I really love that you share the inside baseball of figures training

@Figure_robot Woah how tall is that? It looks like it’s 6’3” tall lol.

the tedious work, is important. neat stuff

Electrician here...can it do ladders?

Excited for Figure to become the first pure humanoid robotics Bluechip in the market.

bootstrapping the recursion more robots -> better robots -> more robots getting off zero is brutal but it looks like we are there now

sims make those stair loops cheap

@Figure_robot Make it skip a stair!

看着只是简单上下楼，实际要连续两小时不摔倒，工程难度应该完全是另一回事

Nice. Now add a ladder

I like how we can see the aura of F.03 on the monitor screen

wen next live stream

Excellent progress, keep it up!

stair loops prove repeatability, but autonomy usually breaks on one scuffed tread or shifted lighting

Impressive, I was not expecting this level of progress in robotics this early in the game.

A first meeting between a human and a humanoid. A simple handshake. A gentle light. A feeling of warmth and comfort. “It feels warm.” “It makes me feel calm.” As technology advances, could kindness, comfort, and human connection become just as important as performance and functionality? This is one proposal for a future where humanoids support people not only with technology, but also with compassion. What kind of relationship would you like to build with a humanoid?

Stairs are brutal