Alphabet launches Intrinsic unit to build software for industrial robots

Alphabet just launched a new company called Intrinsic out of its X tech prototyping unit as an independent entity to develop software tools for industrial robots.

The goal is to make industrial robots easier to use, less costly and more flexible, so they can be used to make products and services primarily for the auto, electronics, and health care industries.The technology has been under development for more than five years at Alphabet, the parent of Google, according to a recent blog by Wendy Tan-White, the new Intrinsic CEO.  Now 50, Tan-White started Moonfruit in the late 90s, a SaaS website builder.

“The surprisingly manual and bespoke process of teaching robots how to do things, which hasn’t changed much over the last few decades, is currently a cap on their potential to help more businesses,” she wrote.  “Specialist programmers can spend hundreds of hours hard coding robots to perform specific jobs, like welding two pieces of metal or gluing together an electronics case.”

Many delicate tasks such as inserting plugs or moving cords “remain unfeasible for robots because they lack the sensors of software needed to understand their physical surroundings,” she added.

“Our team has been exploring how to give industrial robots the ability to sense, learn and automatically make adjustments as they’re completing tasks, so they work in a wider range of settings and applications,” she said. Working with partners, the group has been testing software that relies on automated perception, deep learning, reinforcement learning, motion planning, simulation, and force control.

She included videos showing how they trained a robot in two hours to make a USB connection, a task that would normally take hundreds of hours to program.  Another test shows how Intrinsic orchestrated robot arms to assemble a simple piece of furniture.  “None of this is realistic or affordable to automate today,” she noted.  Another example cited by Alphabet engineers is the inability of a robot to properly tighten a screw.

Ultimately, Intrinsic could reduce time, cost and complexity for using industrial robots. The timing of the Intrinsic announcement comes amid growing advances in software for industrial robots, even as Industry 4.0 and the Industrial Internet of Things seem to have stalled in many industry segments.  Nvidia and USC researchers recently announced a simulator for robotic cutting that predicts the forces acing on a knife as it slices through fruits and vegetables.

RELATED: Sushi robot gets a fancy sim trainer

Today only 10 countries make 70% of the world’s goods, Tan-White noted.  “This means most things are made far away from their end consumers,” driving up emissions and keeping many countries and businesses from economic opportunities.  U.S. manufacturing is expected to have 2 million unfilled jobs by 2030, according to Deloitte.

In addition to looking for partners, Tan-White said Intrinsic is looking to hire AI experts, software engineers and roboticists.