Report Card for Donald Trump

[pbrower2a]

(01-04-2017, 12:53 PM)SomeGuy Wrote:

(01-04-2017, 12:31 PM)David Horn Wrote:

(01-03-2017, 06:09 PM)SomeGuy Wrote: No, Dave, the time horizon is not my only issue.  Moore's Law is already showing signs of stalling, resource limits/AGW/assorted Limits to Growth stuff WILL begin to impact the economy over the course of the 21st century, global supply chains and capital markets may not survive indefinitely, and improvement in technology alone != THE END OF ALL JOBS FOREVERS <cue appropriate cat emoticon here>.

Unless we develop quantum computing, which I believe we will at some time in the not-distant future, Moore's Law must have an upper limit ... but so what?  We already have plenty of horsepower in the hardware we currently deploy.  As the HW guys always note: they build it and the SW guys waste it. 

So most of the development work is systems level (voice recognition and vision being the two most important).  We already know how to do expert systems, and AI just needs the systems interface to complete the circle.  Since these systems don't have to be general purpose, much of the complexity can be eliminated along with the cost.  In any case, global supply chains are not really an issue.  The HW just needs some silicon, plastic and a bit of gold and silver.  The rest is simpler than a toaster: circuit board material, aluminum for heat dissipation and whatever material for physical structure.  Only the SW requires a lot high level input, and we're already nearing the point where the SW is "adequate" as-is.  The rest is just a version of what we already have.

What makes you think that average run-of-the-mill humans can compete with machines that are cheap to build, almost certainly self-maintaining and available 24/7?  Add to that, specialty machines will be better at what they do than any human, no matter how talented.

Emphasis mine.  Is the SW adequate as is, or does it require more development work?  What's the power consumption like?  How on earth can something be "self-maintaining" and also "simpler than a toaster"?  Is all this stuff going to be manufactured here?  Will it manufacturer itself?  What about liability concerns?  Consumer uptake?  Is machine-grown food today more or less prestigious than low-tech alternatives?  Are people really going to want interact solely with robots all day, for every task?  Are these only going to be software systems or industrial robots, or are they going to be walking around, grooming dogs, carrying groceries?  How much will these things cost?  Is everything going to be free in the glorious communist paradise of progress?  Materials, energy, land, time, etc?

These predictions were being made when you were a child, older even.  It was always just 20 years out.  No, seriously, I have a book on the history of cybernetics as an idea sitting on my bookshelf, I can supply quotes from people you'd heard of talking about the imminent (1960s) end of work.  The gains made between the Wright brother's maiden flight and the Apollo Moon landings were equally impressive in their day, and a lot of wild predictions by serious people were made.  Where are the flying cars?  Jet Packs?  Personal helicopters?  Rocket ships to Mars?  FTL travel?

Look, I am sure amazingly clever things will be done within my lifetime, but all I am hearing right now is a lot of overheated rhetoric recycled from the Space Age coming from people spending too much time reading press releases and not enough time actually thinking things through.

We are creating increasingly-complicated devices so complex that no single person could create them. An inventor like Thomas Edison would be unable to work with the technology of our time as he could with the comparatively-primitive technology of his time (even if it is still useful) because he invented objects that were very basic by modern standards.

Innovations too complicated for manufacturing or use will not reach, let alone remain on, the market. Technologies that devour huge amounts of resources or leave behind dangerous consequences (like highly-toxic effluents) will not make it either. Is there reality beyond the understanding of more than a handful of people? The much-lauded devices of our time are simply miniaturized versions of what we used to have, miniaturization possible due to more intricate, tiny components capable of doing things with more sophistication and leser material. The reader device on which one can download a whole edition of The Brothers Karamazov free of charge has only one constraint on its size -- the willingness and ability of people to read fine print. (I have no idea of how long one will be able to get recorded music free of charge, unless advertiser-supported, on YouTube, but it is remarkable that one can get this on a reader





with an internet connection, and on an object that weighs less than a pack of cigarettes and costs less than a carton of smokes.

Of course it is not that simple: I needed a search engine which someone had developed at considerable effort... and someone had to have means of downloading a cinematic film of the opera Prince Igor. The reader is easy to use, but that reflects some complexity in design to make it work.

Will we demand something more complex? I doubt it. Screen size and keyboard size have the obvious limits of our vision and the size of our fingers -- indelible characteristics of biological reality. Maybe we could add a voice command (as in "Get me the opera Prince Igor") to process the search, which would be good for people whose fingers are either non-existent or unworkable (quadriplegia, Parkinsonism, amputation, recently-broken arm) by attaching a microphone to the processing unit.

The one constraint on computing is that it can't quite imitate human creativity. Can a Mozart piano concerto or a Hokusai painting be reduced to algorithms that suggest how to create something similar in quality yet different enough to be a unique work?  Now that may overpower the ability of people to program the orders to do such.