Not a member yet? Register for full benefits!

Podcast: Hod Lipson: Robots that are "self-aware"

Podcast Source:

View Podcast Online? Yes

Podcast length: 6 minutes 18 Seconds

Podcast Description

Part of a talk from TED 2007. Hod Lipson demonstrates a few of his cool little robots, which have the ability to learn, understand themselves and even self-replicate.

Presenter Biographies

Hod Lipson

To say that Hod Lipson and his team at Cornell build robots is not completely accurate: They may simply set out a pile of virtual robot parts, devise some rules for assembly, and see what the parts build themselves into. They've created robots that decide for themselves how they want to walk; robots that develop a sense of what they look like; even robots that can, through trial and error, construct other robots just like themselves.

Working across disciplines - physics, computer science, math, biology and several flavours of engineer - the team studies techniques for self-assembly and evolution that have great implications for fields such as micro-manufacturing - allowing tiny pieces to assemble themselves at scales heretofore impossible - and extreme custom manufacturing (in other words, 3-D printers for the home).

Transcript Available? No

Audio file available? No

Podcast Download? Yes

20.2 MB

Podcast viewing notes

This talk starts out on a humorous note: "So, where are the robots? We have been told for 40 years now, they are coming soon, very soon."

So, he and his team decided to call upon the 'mother of all designers' and see if it was possible to evolve a complex robot, rather than just building them outright.

A huge variety of machines were created within a virtual environment, which learnt to move, and moved in completely different ways. The best designs - those which showed the most promise, moved forwards and were altered time and time again.

Eventually the best designs were physically created, scurrying like insects. This design was not intentional, its just how the best movers evolved from the process.

All the physical robots, same as the virtual, learn to move autonomously. None of them start out knowing what they look like, or how they are configured. They have to learn this for themselves, the same as a baby does.

"In the absence of any reward, the intrinsic reward is self-replication". An astute statement garnered from watching a couple of thousand simple robotic blocks in a room, all using this no idea what they look like, teach themselves to move AI, and allowing them to copulate and bond. Over time, lifelike patterns start to emerge, then replicate.

Additional Research Links

Hod Lipson

Cornell Team

Cornell Robotics Outreach

Staff Comments


Untitled Document .