Tutor Intelligence Inc., a provider of artificial intelligence-powered warehouse robot workers, has launched its second-generation Cassie and Sonny robots, running on the company’s robot foundation models.
Sonny was initially developed as the company’s dual-armed semi-humanoid robot and introduced alongside Tutor’s Ti0 4.5 billion parameter Vision Language Action model. Since then, it has evolved from a research platform into a fully deployable robot capable of autonomous manipulation and mobility in warehouses.
Cassie, the company’s large-format, single-arm robot, was expanded for mobility in warehouses and factories, adding pallet movement and greater physical intelligence. It looks like a giant arm atop a large box with wheels.
“It’s really important for us not only that we build these general robots, but they’re immediately practically useful,” co-founder and Chief Executive Josh Gruenstein told SiliconANGLE in an interview. “We’re in this unique position of building robots that are both generally capable and immediately useful.”
Tutor looked at an industry where humanoid robots are currently “sexy,” but humans already do that fundamental work just fine. Cassie and Sonny represent two separate answers to the same question: What shape can a general robot take if it’s designed around the work rather than around looking like a person?
Cassie provides muscle for moving heavy material in a broad, big box-laden pallet warehouse setting and Sonny works amid shelves where arms and hands make sense for picking up small objects and moving them between bins.
“Sonny is kind of this co-design with Cassie of what is something that is human-like as a robot, but, you know, not designed just to mimic a human, but to do a class of work that humans do,” Gruenstein said.
Cassie in particular can lift 50 pounds and tug thousands of pounds, which Gruenstein described as something of a “loch ness” of a robot moving between pallets doing work. The first-generation of Cassie worked from a fixed position, but is now mobile and can move across warehouse floors.
From the classroom to the factory floor
Back in December, Tutor raised $34 million in Series A funding; in that time, the company built what is essentially a “classroom” for its AI models to run across its Sonny robots in what it calls Data Factory 1: a 100 “robot factory” where Sonny robots “learn.”
“In robotics, we don’t have like an internet of training data to get the training data to train AI models,” Gruenstein explained.
Robotics has a mild problem. AI models depend on vast amounts of data for pre-training; in robotics, that data is visual and kinetic telemetry that shows them how to do tasks such as picking up objects, what objects look like, where to place them, how to move and so on. Much of this can be done through training, or, as it’s called, “tutoring,” via experts who show the robots how it is done in real settings.
According to Gruenstein, Cassie doesn’t need to reason about everything a pair of hands might encounter. Its physical vocabulary is comparatively small: Identify an object, grasp it, move it and place it. Sonny’s vocabulary is much larger, so Tutor built Ti0, the company’s powerful AI model, and DF1 to attack the corresponding data problem.
“We have this team of tutors staffed internationally that can kind of remote control a robot and teach it how to do a new task for the first time,” Gruenstein said.
After that, the 100 robots in the classroom go through the motions to fine-tune behaviors, smooth out any kinks and prepare the robotic models for what might happen in the field. This is otherwise known as post-training. If an individual robot does a good job, that student’s actions get reinforced in the AI model with a thumbs up, a positive reward; if it does poorly, those actions get a thumbs down. This helps the model learn how real-world behaviors interact with actual hardware.
“We can get robots that recursively self-improve and are superhuman in their performance relative to people,” Gruenstein explained.
