Why toddlers are smarter than computers | Gary Marcus | TEDxCERN

Why toddlers are smarter than computers | Gary Marcus | TEDxCERN

Advancements in artificial intelligence are changing how we analyze and process information. But these advances fall short when compared with the ingenuity and creativity of an average three-year old. In his talk, psychologist and neuroscientist Gary Marcus compares advancements in AI to the resilience of the human brain on tasks such as common sense and real-time evaluations.

“Gary F. Marcus wants to build a human mind from scratch. His research at New York University combines psychology, linguistics, and molecular biology to map the inner workings of the brain and deconstruct common sense scientifically.
“No modern machine can match the ingenuity of any three-year-old,” Marcus said. “A deep understanding of the human mind will radically transform society.”
In addition to teaching, Marcus is a best-selling author and founder of the company Geometric Intelligence, which is redefining the boundaries of machine learning. By applying a deep understanding of the human mind to artificial intelligence, Marcus hopes to revolutionize the capabilities of modern technology and radically transform society.”

This talk was given at a TEDx event using the TED conference format but independently organized by a local community. Learn more at http://ted.com/tedx


  1. Justin Baker on September 21, 2019 at 8:30 pm

    Very sober analysis.

  2. Yao Chen on September 21, 2019 at 8:32 pm

    This video was recorded just two years ago, but seams so ancient in the field of AI.

  3. Andrew Craton on September 21, 2019 at 8:41 pm

    Shocking that researchers are JUST NOW figuring this out. OF COURSE, natural language AI doesn’t understand anything. It doesn’t understand "by design". It has no experiences, so it simply can’t understand in a way that things that experience do.

  4. RSH Productions on September 21, 2019 at 8:45 pm

    The problem with AI is a result of the fundamental way computers process data you have an input say a math problem the computer takes that data and begins to go through possible solutions via process of elimination because processors only understand 1 and 0 yes or no on or off. So until a computer can understand more than yes or no AI wont really be feasible even quantum computing may not be enough for true AI.

  5. Brando Miranda on September 21, 2019 at 8:48 pm

    isn’t this what Deep Mind sort of is?

    nice, talk. 🙂

  6. vizeet srivastava on September 21, 2019 at 8:53 pm

    True…computers are way behind compared to humans but that does not change the fact that most jobs don’t require human capability, they require capability which deep learning algorithms already provide and big data and sensors and rapid simulations can help them train quickly.

  7. Javier Rey on September 21, 2019 at 9:06 pm

    "but it really matters when it comes to things like medicine, we want the AI to really do it right" – didn’t his AI company get bought out by Uber to build self driving car tech? :/