Skip to Content
Artificial intelligence

VIDEO: Boston Dynamics’ Backflipping Robot Is an Astounding Advance

November 17, 2017

What a difference two years make. If you followed the DARPA Challenge back in 2015, you’ll know that humanoid robots have a track record of falling over—a lot. Fast forward to today, however, and the notorious robot maker Boston Dynamics has gotten its two-legged Atlas humanoid to do backflips. Just watch it. It’s literally, for me at least, jaw-dropping.

Atlas has been slowly finding its feet for years: learning how to roam snowy hillsides, traverse rocky paths, and ensure that, if it does fall, it at least does so safely. Still, there have been detractors along the way: a clip of the robot getting confused and collapsing went viral earlier this year. “Judging from the progress made by arguably the most advanced humanoid robots, if Terminator does show up, it won’t even be able to stack shelves without falling over,” wrote the Guardian at the time.

Um, nope. Anyone who’s tried to do a backflip will know that it’s not easy, especially the landing. It requires dexterity, poise, and balance—all things that have typically been lacking in humanoid robots. But Boston Dynamics has shown that, while it might not be easy and certainly seemed a distant hope two years ago, it’s now perfectly possible for a robot to display such skills.

That has profound implications for what robots might be able to do in the future, and it makes the prospect of, say, reliable rescue robots a very real possibility. It also means that you can scratch that Guardian quote: if Terminator does show up, he’ll school you at parkour.

Deep Dive

Artificial intelligence

What does GPT-3 “know” about me? 

Large language models are trained on troves of personal data hoovered from the internet. So I wanted to know: What does it have on me?

An AI that can design new proteins could help unlock new cures and materials 

The machine-learning tool could help researchers discover entirely new proteins not yet known to science.

DeepMind’s new chatbot uses Google searches plus humans to give better answers

The lab trained a chatbot to learn from human feedback and search the internet for information to support its claims.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.