1/ 2 mm.
That’s the median TCP tracking error we’re now achieving across our latest collection activities with our in-house SLAM stack.
This clip was captured inside a 300-person manufacturing site in Delhi, India.
3/ Our stack combines synchronized stereo egocentric vision + wrist-mounted UMI + SLAM to recover precise 6-DoF human manipulation trajectories in a common world frame.
At 2 mm median TCP error, we’ve made high-fidelity action capture deployable at real-world scale.
4/ Now we want to push it further.
If you’re exceptional at SLAM, tracking, sensing, or robotics hardware, we’re hiring:
careers@neocambrian.ai
And if you want access to the data these rigs produce, we’d love to chat:
hello@neocambrian.ai
India is a goldmine of diverse data. People looking down on data businesses don’t take the bitter lesson and history of DL seriously enough. Creating imagenet as one of the first hills to climb was as key a contribution as running NNs on GPUs.
The “scale” of these businesses in *each* domain - text, vision, driving, speech has been crazy and embodied intelligence is both the most starved and likely the hardest.
Today, we announce @NeocambrianAI
The future of AI will be physical. But Physical AI has no internet scale dataset to learn from.
We’re building the data foundation of Physical AI - a high fidelity, pre training scale database of Human Action.
From India. For the world.
Today, we announce @NeocambrianAI
The future of AI will be physical. But Physical AI has no internet scale dataset to learn from.
We’re building the data foundation of Physical AI - a high fidelity, pre training scale database of Human Action.
From India. For the world.
Today, we announce @NeocambrianAI
The future of AI will be physical. But Physical AI has no internet scale dataset to learn from.
We’re building the data foundation of Physical AI - a high fidelity, pre training scale database of Human Action.
From India. For the world.
Today, we announce @NeocambrianAI
The future of AI will be physical. But Physical AI has no internet scale dataset to learn from.
We’re building the data foundation of Physical AI - a high fidelity, pre training scale database of Human Action.
From India. For the world.