[extropy-chat] 22nd Century Electronics

[Adrian Tymes]

--- "nvitamore at austin.rr.com" <nvitamore at austin.rr.com> wrote:
> Does anyone have an idea of what electronic (manufacturing) might be
> in the
> next 100 years?  

100 years is way too far to safely predict.  That said, extrapolating a
few current trends (in ways that others before me have extrapolated),
here are some guesses:

* DIY is already on the rise, with more and more prefabricated
  components for the user to assemble, and in extreme cases personal
  fabrication machines to make those components.  The range of
  materials they can process is presently quite limited (especially
  WRT metals), but it seems safe to say that personal fabrication
  machines will be able to tackle these within a few decades, and will
  be at least as common as power tools are today.

* At some point, probably before 100 years, a sufficient number of
  these machines will be deployed in some urban area (city, or maybe
  more likely city-equivalents that have yet to be built, like
  high-population space colonies or city-in-a-building arcologies, but
  only if one of those trends takes off) as to make it viable to
  install feedstock pipes as a utility, just like water and
  electricity.  (This would also take care of recycling the scraps from
  subtractive milling machines, which otherwise might become a minor
  "industrial" pollution problem.)

* This also means you'd see the current "open source vs. proprietary"
  software battle in hardware.  Of course, we are already seeing the
  start of that today, with services that let you write your own chip,
  upload the design, and pay a fee, then they print a few copies of
  your chip and ship them to you.  See http://www.opencores.com/ and
  others.  Moving the chip writers into homes would just make that more
  prominent.  Also, cheap custom chip fabrication has to work at
  coarser resolutions than high-end plants like those producing Intel's
  and AMD's wares, but even if Moore's Law slows down as we near atomic
  scale electronics, it's likely that personal manufacturing facilities
  will catch up to atomic scale in 100 years.

* Of course, Intel and AMD probably won't passively sit around as the
  public catches up.  See the already heavy R&D in quantum computing,
  DNA computing, and all-optical computing.  It may be an open question
  which one will dominate (or if something else will), but in 100
  years, top-end supercomputers (and maybe personal computers) will
  likely not rely on electrons as we now use them.  (I'd give slightly
  better odds to all-optical, at least in some form, for personal
  computing since the other two can speed up parallel processing but
  not - at least for poorly-parallelizable problems - serial
  processing.  Also, all-optical doesn't need the special cooling or
  nutrients of the other two when doing nothing, so it'd be easier to
  maintain.)

* Also expect to see more and more of the industry shift to robotics
  for manufacturing, even aside from the above-mentioned personal
  fabrication machines (which need robots because the whole point is
  sellable trained labor, and even in 100 years it seems unlikely that
  society would accept shrink-wrapped humans).  Once you get the
  manufacturing process down, you simply take the movements you
  guided the robot arms through during R&D, possibly refine them, then
  copy and paste to the factory floor.  Also, as robots get cheaper, it
  becomes more feasable to flood a manufacturing site with robots, send
  them through dangerous (but cheap) processes, and if (when) they get
  damaged...eh, they're cheap, buy another.  (Which is very much not
  the situation today: human labor is often much cheaper than robots,
  especially where minimal skill is required.)