According to a recent posting on Slashdot, MIT researchers have succesfully implemented a system which mimics how the brain performs visual recognition, at least in a limited context. Interestingly, they were successful by building capabilities which closely matched the way the brain is believed to accomplish this function itself, rather than attempting more 'mechanical' approaches.

These capabilities and mechanisms are consistent with those suggested in How the Mind Works and The Singularity is Near. Of course, the natural course of evolution suggested by these works, and impacts to technologies and systems which might use those capabilities, can escalate rapidly.

Though it will likely be years before such ideas are commercialized, it would seem like the applications of such technologies are enormous. Consider this: once voice recognition systems with sufficient functionality began to be available, there was a rapid decrease in the following kinds of jobs:

• medical transcription
• telephone operators
• secretaries
• catalog order taking people

Lots of jobs were lost (if they hadn't already gone overseas). There are many more jobs, though, that depend upon visual recognition, including:

• mechanics
• surgeons
• transportation operators (pilots, drivers, etc)

These are higher paid positions, which mean that the value proposition for replacing them will be easier to make from the 'cost of replaced labor' perspective (though the corresponding demanding requirements for accuracy could work to counter this, at least initially). When machines can see, maneuver, and think, we'll find robots doing far more than just vacuuming.

In parallel, impressive progress is also underway in bionic eyes. Who will get there first - robots with human-like brains, or humans with robot-like vision?