George Rebane

A few days ago IBM published a very short paper titled simply 10¹⁴.  It summarized the first successful testing of TrueNorth, IBM’s cognitive computing testbed funded under DARPA’s SyNAPSE.  That program “calls for developing electronic neuromorphic (brain-simulation) machine technology that scales to biological levels, using a cognitive computing architecture with 10¹⁰ neurons (10 billion) and 10¹⁴ synapses (100 trillion, based on estimates of the number of synapses in the human brain) to develop electronic neuromorphic machine technology that scales to biological levels.”

It so happens that the human brain contains about 10¹⁰ neurons and 10¹⁴ synapses in a system of relatively few highly organized and repetitive pattern recognition structures arranged into hierarchies.  This is a major advance in the current fanout of intelligent machines that seek to replicate the functioning of the human brain.  Right now TrueNorth runs only about a 1,500 times slower than the human brain.  But that will soon change.

I am currently in the middle of Ray Kurzweil’s How to Create a Mind (just published) wherein he goes into great detail on the type of structure and capabilities of a neuromorphic machine like TrueNorth.  Therefore I was happy to see this concurrent announcement of the next step toward Singularity from IBM, the company that also brought us Watson (here) and is now marketing that into several application areas beginning with medicine.

These advances need to be viewed in the context of today’s accelerating technology, and its disturbing impact on the workforces of developed countries (more here).  Their greater effect on humanity is already very difficult to predict beyond giving rise to the obvious redundancies of mediocre human talents.