AI meta-point-missing
Charles Rubin reviews a book on tech dreams and myths in ancient times.

As we move from chapter to chapter, we accumulate fascinating information, but the book does not obviously develop an argument about the lessons we might learn from it. Mayor hopes that “rereading those ancient stories might enrich today’s discussions of robotics, driverless cars, biotechnology, AI, machine learning, and other innovations.” And indeed, almost by definition, to find out that thinking about these matters goes back far longer than we might have expected is enriching.

Without trying to read the original, it sounds like both the author and reviewer are missing a basic distinction. (I could be wrong here because the reviewer could be wrong about the original.)

Patient machines have been around for thousands of years. Negative feedback, the essential difference between living and non-living, has been part of devices for thousands of years. These devices, from drain traps to toilets to thermostats to automatic lighthouses, operate independently like living things, running their own simple lives without human control.

In ancient times full-fledged automata, literal robots, were either a dream or a scam. The famous automatic chess-player was a scam, containing a midget who operated the robot arm.

In modern times full-fledged automata, literal robots, are STILL either a dream or a scam.

Autonomous cars will never work. Factory automation has been around since Jacquard, and STILL isn't much beyond Jacquard. Some routine operations can be automated and have been automated for a long time. The peak point was reached 50 years ago. Idiots like Elon keep trying to break through the peak, and keep failing.

= = = = =

The reviewer raises a better question that doesn't have a simple answer.

Mayor notes, but hardly lingers to investigate, that in the mythic record there is frequently a link between robot-like beings (or other innovations) and tyrants. Shouldn’t this strike us as odd, given that in the modern world successful technical innovations seem more associated with the free world than with tyranny?

I don't think there's a correlation either way. Many important inventions were developed by armies or for armies, and strong empires tend to have bigger and better-funded armies. In that class of innovation, empires are the leader.

But other inventions come from a variety of needs, which exist in all sorts of governing systems.

Example from my most recent 'tutorial': The Optophone was developed from a desire to help blind people. Its later rejuvenation happened with military funding, attempting to help blinded veterans.

The only real constant is a conjunction of two factors: available funding and perceived need.

Military tyrannies like USA have plenty of available funding for the needs of Deepstate, but very little for tinkering.

Small countries with looser governments may leave more room for tinkerers to work on perceived needs.... BUT the best tinkerers in the modern world are in socialist countries like Cuba and Persia, where our blockades create the need for many inventions that would otherwise be bought from China.

Labels: $TSLAQ, AI point-missing, Asked and badly answered, Patient things

  ¶ 1:59 PM