Two Views of the Post PC World - Automata Processor and TOMI Celeste, Part 3: Ray Kurzweil's Crazy Dream

-July 01, 2014

In Part I and Part II we looked at the design and development systems of two post-PC architectures: the Micron Technology pattern-matching state machine Automata Processor and the Venray Technology 8-core/64-bit/2Gbit processor in memory TOMI Celeste.
For a look into post-PC applications of these architectures we've chosen as our guide, Ray Kurzweil.
"To transcend means to go beyond, but this need not compel us to an ornate dualist view that regards transcendent levels of reality (e.g., the spiritual level) to be not of this world." -- Ray Kurzweil

Ray Kurzweil may be barking crazily at the moon, but questions of sanity aside, Kurzweil sees the future in 3D 4K HighDef clarity.
Kurzweil is not the only mad scientist racing headlong into this brave new post-PC world.  Palm founder Jeff Hawkins has spent a decade developing his "Hierarchical Temporal Memory" model of the human cortex.  Similarly in Hong Kong, Ben Goertzel has been flogging his "Probabilistic Logic Network" to implement artificial general intelligence.
Ray distinguishes himself from the others by laying out a detailed roadmap to the future including many product application mileposts along the way in his trilogy:
Not only does Kurzweil see the future, he has a history of making it happen as he has done with optical character recognition, music synthesis, speech recognition, and stock trading.  In Ray's head the world is nothing more than a collection of patterns: tempo of a musical piece, pixels forming characters, phonemes in spoken words, or prices of equity trades.
His latest infatuation is artificial intelligence and robotics.  With his financial security assured and at the age where most are receiving Social Security, Ray already has his brass ring.  Now he is playing for the history books.
He's agreed to become Google's Director of Engineering.

In that role he commands one of the world's largest and most sophisticated AI labs and design teams and directs an acquisition budget that for all practical purposes is unlimited.
At Google he can tinker with patterns to his heart's content.  Of course the ultimate in pattern management is far more than a voice operated word processor or a cell phone that identifies its user.
Christopher Bishop succinctly described the relationship between pattern recognition and machine learning,
From his pattern-centric view Ray sees a world where learning machines successively enhance human senses and abilities and finally supersede them.  Eventually at some point in the future man and machine merge.
In the decades before that does or does not happen, there is plenty of money to be made on more prosaic artificial intelligent applications, which is of course why Google drafted him.
We'll look at two brand new product applications from Ray's vision, but first some historical context and a quick review of Micron’s Automata and Venray Technology’s TOMI Celeste.
Quick History of the PC World

The PC world roughly began in 1975 with the introduction of the MITS Altair 8800, based on INTEL's 1MHz 8080 8-bit microprocessor.
(Historical note: Purists will observe that the 8080's external clock was 2 MHz.  However the clock was immediately divided by 2 internally.  Some have argued that INTEL used the divide by 2 trick to imply that the 8080 was superior to the very similarly performing 1 MHz Motorola 6800.)
For the next 3 decades, INTEL and the PC industry rode Moore's Law and Denard Scaling to rapidly increase performance and lower cost.  Clock speed increased over 3,000X until INTEL ran head first into the Power Wall with the 2004 Pentium IV at 3.73 GHz.

Loading comments...

Write a Comment

To comment please Log In