From Player Piano to Master Chef: Memo Learns Like a Human

For decades, robots have behaved like player pianos. Precise, repeatable, and impressive in controlled environments, but ultimately limited to performing only the exact actions they were programmed to execute. Give them a new object, a slightly different setup, or a novel task, and they quickly fall apart.

That approach is reaching its limit.

At Sunday Robotics, the team behind Memo is taking a fundamentally different path. Instead of treating robots as rigid machines that follow scripts, Memo is being trained more like a human apprentice.

Think less player piano, more Master Chef.

Memo has observed hundreds of people interacting with objects across countless real-world environments. Different homes. Different kitchens. Different hands, tools, shapes, textures, and habits. From this diverse dataset, Memo does not memorize instructions. It builds intuition.

That intuition allows Memo to approach entirely new objects with a sense of how they might be grasped, lifted, or manipulated, even if it has never seen that exact object before. A mug with an unfamiliar handle. A toy on the floor. A tool left on a counter. Memo reasons about the situation instead of freezing.

This shift matters because real homes are messy, inconsistent, and endlessly varied. No two households are the same. Robots that rely on rigid programming will never scale into everyday life. Robots that learn from broad human experience can.

Memo represents a small but meaningful step toward truly helpful robots. Not machines that demand perfect conditions, but assistants that adapt, learn, and quietly fit into the background of daily life.

When robots gain intuition, they stop feeling like machines and start feeling useful.