This autonomous bicycle shows China’s rising expertise in AI chips
It might not look like much, but this wobbly self-driving bicycle is a symbol of growing Chinese expertise in advanced chip design.
Look, no hands: The bike not only balances itself but steers itself around obstacles and even responds to simple voice commands. But it’s the brains behind the bike that matter. It uses a new kind of computer chip, called Tianjic, that was developed by Luping Shi and colleagues at Tsinghua University, a top academic institution in Beijing.
Two in one: The Tianjic chip features a hybrid design that seeks to bring together two different architectural approaches to computing: a conventional, von Neumann design and a neurologically inspired one. The two architectures are used in cooperation to run artificial neural networks for obstacle detection, motor and balance control, and voice recognition, as well as conventional software.
AI’s future? In a paper outlining the chip and the bicycle, published in the journal Nature today, the researchers suggest that such a hybrid architecture could be crucial for the future of artificial intelligence, perhaps even providing a route toward more general forms of AI. That’s a bit bold, given how far we are from AGI, but Tianjic does show the growing value of new chip designs optimized for running AI algorithms.
Made in China: The chip also hints at the progress China is making in developing its own chip design capabilities. As outlined in this feature article, China has long struggled to build its own chip industry, a major weakness in its technological capabilities that have been exploited in the ongoing trade war with the US. But while manufacturing the most advanced computer chips remains out of reach, Chinese researchers are showing they can make specialized AI chips as well as anyone.
Deep Dive
Artificial intelligence
Large language models can do jaw-dropping things. But nobody knows exactly why.
And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.
Google DeepMind’s new generative model makes Super Mario–like games from scratch
Genie learns how to control games by watching hours and hours of video. It could help train next-gen robots too.
What’s next for generative video
OpenAI's Sora has raised the bar for AI moviemaking. Here are four things to bear in mind as we wrap our heads around what's coming.
Stay connected
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.