Last month in Beijing a robot named Lightning shattered the half-marathon record.
Beaten by nearly seven minutes? It ran circles around the current human best.
This isn’t an isolated weirdness. It is the latest blip in a string of AI milestones that make you stop scrolling. People are actually wondering if robots are about to step out of warehouses and into our kitchens. Just like chatbots did.
The place making this happen?
China.
They are throwing serious money at this problem. Over the next twenty years the government has pledged more than £100 billion for robotics. That is not pocket change. It is a massive industrial shift.
We got Ian Sample talking to the people actually in the thick of it. He chatted with Amy Hawkins. She is the senior China correspondent for The Guardian so she knows where the money is going. He also talked to Nathan Lepora.
Who is Lepora?
He’s a professor of robotics and AI at Bristol. He figures out how to give metal fingers human-like dexterity. It turns out making a robot run is easy compared to making it tie its shoes or clean your house without breaking the vase.
What needs to break before robots move in
It is not just about speed.
Lightning won on pavement. But can it handle the chaos of a living room?
The goal isn’t just to watch them race. It’s about utility. Cleaning. Weeding. Doing the chores we all pretend we are too tired for.
Nathan Lepora argues that we are hitting a wall of complexity. Hardware is getting cheaper. Software is getting smarter. But the physical world is messy. Things slide. Gravity doesn’t care about your algorithm.
Amy Hawkins notes that the scale of China’s investment changes the timeline. If a country pours that much cash into R&D you start seeing results in five years. Maybe three.
Who is going to hire the first thousand units?
Probably not you. Yet.
Factories first. Warehouses always. Then maybe construction. The jump from “holding a fixed pipe” to “picking a tomato” is massive. It requires a brain that can adapt in milliseconds.
The dexterity gap
This is the bottleneck.
Bristol’s researchers are looking at how muscles work. Not just joints. Tension. Feedback loops. Real humans adjust pressure constantly. A robot slams down unless we program it otherwise.
We are
