AI Just Learned to Grab a Wrench

For a few years the AI story lived inside a browser tab. Chatbots, copilots, text in a box. That story is now growing hands and walking into warehouses. Two moves this month make it real: a humanoid hand built to match yours, and a wearable badge that maps physical space as workers move through it. Let me catch you up on what changed and what to do with it.

The hand was always the hard part

The thing blocking robots from real work was never the brain. It was the fingers. Human hands are precise and delicate in ways machines could not copy. So the field spent decades building around the problem instead of solving it. 1X put it plainly: "For seventy years, robotics worked around the hand problem. The humanoid bet is the reverse."

Their Neo hand has 25 joints and a tendon system, with the motors moved into the forearm the way yours are. It can pick a single screw out of a wallet, spin in a lightbulb, use a screwdriver, and handle origami without tearing it. Tesla made the same forearm move on its latest Optimus hand. Two companies, same design choice, same bet.

Made in the thousands, not the lab

Demos are cheap. Scale is the tell. 1X says it can build 10,000 of these hands this year, in house, and the first Neo units ship to households this year. That is a production number, not a research slide.

The hands also carry sensors across the fingertips and surface, so the robot feels pressure and contact. It can tell when something is about to slip. Worth noting the human reaction: writer Stan Schroeder called the video "very, very unsettling," especially the fingers bending backward. If your product touches customers or workers directly, that discomfort is a design problem you will own, not a footnote.

The badge that watches the aisle

While the robots learn to grip, another approach skips the robot entirely and puts AI on the worker. Augmodo straps a dual-camera badge onto store staff. As they walk the aisles, it builds a live 3D map of the shelves. No new headcount, no robot, just spatial AI riding along.

The numbers moved fast. Augmodo raised $21 million at a $350 million valuation, grew revenue 10x in a year, and now maps 186 million square feet a month, aiming for a billion by year's end. CEO Ross Finman said he was not even looking to raise. Customers pulled him into warehouses, factories, and hospitals on their own.

The 80 percent nobody built for

Here is the framing that should stick with you. "Everyone's focused on the 20% of the workforce that's knowledge work," Finman said, "and we're focused on the 80% that's physical work." Your AI roadmap probably lives in that 20 percent: writing, coding, summarizing, support tickets.

What both companies show is that the same models now work on the shop floor. "Someone grabbing a wrench at an automotive factory isn't that different from someone grabbing a Cheerios box," Finman said. The algorithms carry across. That means the tools built for retail can walk into your operations without a rebuild.

The deep cut

The cheaper of these two paths is winning first, and that changes what you plan for. A humanoid robot is a big capital bet years out. A badge lighter than an iPhone that also does walkie-talkie, panic button, and digital ID is a purchase order your ops lead could sign this quarter. Finman turned it into one device instead of five, sold near cost, with features layered on top. That is the shape of what shows up at your door first. So do not scenario-plan for robots marching in. Plan for a wearable that already maps your floor and starts feeding data nobody agreed on how to use. Decide now who owns that data, what it can measure about your workers, and where you draw the line. Bring that to your next review before a vendor brings it to you.

Three questions for your team

• Which of our physical workflows, inventory, maintenance, training, could a wearable like Augmodo's map next quarter, and who would own the data it collects?
• If a tool built for retail already works on a factory floor, are we still scoping AI only for our knowledge workers?
• When machines start handling delicate work near our customers or staff, who on our team owns the "unsettling" problem before it ships?