Terminator building blocks

Implementation of SkyNet and the Terminator are a little behind the original schedule but progress is still being made:

“The team actually started out by building a retina and they came to me and said: ‘Look, it responds to these optical illusions the same way a human does. They put another layer of cells behind that it started to find features, They put another layer, it started to find corners or oriented lines or something, another layer, it started to find patterns,” Jacobs said.

“Today it tracks objects. It’s actually not programmed, it’s taught.”

Jacobs laughed at the silence in the room, conceding he evokes images of “The Terminator.”

SkyNet building blocks are falling into place as well:

The long term future belongs to optical interconnects, low power processors and new kinds of memory architectures, said Prith Banerjee, director of HP Labs in a DesignCon keynote here.

Banerjee described the path to a terabyte/second optical bus as one step toward its vision of future systems architectures. Engineers need to embrace the new technologies to deal with the coming flood of digital data, he said.

“By 2020 your end customer will be living in a world where people access 50 zettabytes of data from 30 billion cellphones and 1.3 trillion sensors–and all that data will have to be analyzed by computer architectures you have to design,” he told a packed audience here.

Share

3 thoughts on “Terminator building blocks

  1. And despite all that the #1 apps will still be Minesweeper and Solitaire and most Internet traffic will be porn.

    Not saying that’s a bad thing.

  2. These kinds of things intrigue me.

    Many years ago, I came across analog robotics in a magazine–you can find information on these things searching for “BEAM robotics”–that are built using schmitt-trigger Nor gates wired up to be oscillators. These tiny robots can learn to walk, be territorial, and cooperate against a common enemy, with a minimum of circuitry. Like the digital stuff described in the article, these devices are largely grown rather than programmed.

    Then again, apparently so is Common Lisp.

    I wish I had time to explore these things!

  3. When I was in college, my design team shared lab space with a professor who researched self-replicating machines, robot evolution and self-taught robots. I’ve always had a sense of pride that, if the machines ever rise up and become our masters, I witnessed the very origins of our demise.

Comments are closed.