Google’s DeepMind AI Teaches Itself to Walk, and the Results Are Kooky, No Wait, Chilling

In 2014, Google acquired Deep­Mind, a com­pa­ny which soon made news when its arti­fi­cial intel­li­gence soft­ware defeat­ed the world’s best play­er of the Chi­nese strat­e­gy game, Go. What’s Deep­Mind up to these days? More ele­men­tal things–like teach­ing itself to walk. Above, watch what hap­pens when, on the fly, Deep­Mind’s AI learns to walk, run, jump, and climb. Sure, it all seems a lit­tle kooky–until you real­ize that if Deep­Mind’s AI can learn to walk in hours, it can take your job in a mat­ter of years.

Watch a primer explain­ing how Deep­Mind works here. And find more AI resources in the Relat­eds below.

via Twist­ed Sifter

Relat­ed Con­tent:

Arti­fi­cial Intel­li­gence: A Free Online Course from MIT

Experts Pre­dict When Arti­fi­cial Intel­li­gence Will Take Our Jobs: From Writ­ing Essays, Books & Songs, to Per­form­ing Surgery and Dri­ving Trucks

Arti­fi­cial Intel­li­gence Pro­gram Tries to Write a Bea­t­les Song: Lis­ten to “Daddy’s Car”

Two Arti­fi­cial Intel­li­gence Chat­bots Talk to Each Oth­er & Get Into a Deep Philo­soph­i­cal Con­ver­sa­tion


by | Permalink | Comments (1) |

Sup­port Open Cul­ture

We’re hop­ing to rely on our loy­al read­ers rather than errat­ic ads. To sup­port Open Cul­ture’s edu­ca­tion­al mis­sion, please con­sid­er mak­ing a dona­tion. We accept Pay­Pal, Ven­mo (@openculture), Patre­on and Cryp­to! Please find all options here. We thank you!


Comments (1)
You can skip to the end and leave a response. Pinging is currently not allowed.
  • Remus Simion says:

    The big dif­fer­ence between what the AI learned to do and what humans do is that the AI was not giv­en an ener­gy expen­di­ture law of some sort. Humans walk dif­fer­ent­ly because they learned to also use ener­gy more effi­cient­ly while work­ing. Would be inter­est­ing to have include such laws as well and see what the AI does :)

Leave a Reply

Quantcast
Open Culture was founded by Dan Colman.