Загрузка видео...
Не удалось загрузить видео
Figure just crossed 86,000 weekly active users uploading data We're building the largest and most diverse robotics dataset in the world The website has a live map viewer showing each pixel of data as it's uploaded around the world
86,816 просмотров • 22 часов назад •via X (Twitter)
Комментарии: 33

You can check out the live map on the site:

How much of the data is lewd

Impressive to see how much Brazil is involved in the data sharing. Hope this sharing is also compensated as one of the first markets outside of US to get the robots.

Accept my application already bro 🥹

Is this real ? WTF is going on in Europe ?

Lots of data from war robots in Ukraine, i see.

Europe is dark. Why is this?

Honest question, nothing in China!? :D

Surprised about Ukraine. Why are they such an outlier in Europe?

I applied from India ,but you did not Approved my application.

every dot a person filming themselves stacking plates, this is the future they promised us

We applied from Turkey, when will you return?

86k people basically doing the data collection job for free is a pretty nasty moat for a robotics company.

Hey Brett, Love your work! I would love to have a chat, just left you a message.

Brett, sign me up and let’s light up the UK on that map!

Why is South Africa so quiet? Let me help you from South Africa 💪

@adcock_brett Stay-at-home Dad here, Ready and willing to help out! 🙏😁

When available in Romania? I want to record and send data. thanks

86k contributors is a massive jump from where this started, the sheer scale of the edge-case variety you're pulling in is gonna make the simulation-to-reality gap much easier to close than people realize

86k people uploading for free is the smartest dataset play in robotics

Italy?

The geographic coverage is useful, but the measurement question is whether the rare viewpoints survive review rather than merely upload. To my knowledge the hard part is not public clips — it is non-public, multi-country real-world video with the messy lighting and viewpoints training sets usually scrub out. We collect that kind of footage if a parallel would help.

There is more activity in the Strait of Hormuz than in Europe

The live map shows geographic reach, but diversity needs task and environment coverage too. I’d publish a breakdown by object type, motion, lighting, failure case, and operator skill so dataset growth maps to capability growth.

What are you guys doing to make sure Figure bot will surpass Tesla Optimus capabilities?

86k WAU is impressive but the dataset value depends entirely on how much of it is labeled. raw pixels don't train robots, annotations do.

86,000 weekly uploaders is a strong scale signal; the benchmark value will depend on exposing task coverage and mid-trajectory failure rates, not just the map of uploads.

👍🏻👍🏻👍🏻💪🏻💪🏻💪🏻

uhmmm the bottleneck was never how much video exists it's how much of it has the right stuff in the frame

Send figure to each of those countries to do tasks on regular houses

Hum, I'm pretty sure Tesla will find a way to synthetically create this data. 100% Elon will not be paying people to wear these devices.

Do you have permission from Bernie and Ro to proceed

When u r launching figure robot?

