Aha, the thread about a time-table for the AI "singularity" moves on to
morphic resonance - my favorite counter-argument to the "bird brain"
stance... which posits that the current state of AI is far from
human-like. It is closer than many of us think with only a few improvements.
Morphic resonance is a natural process of self-organizing systems which
"inherit" both memory and heuristics from previous similar systems.
Think about the "child prodigy" for example. Computers are not exactly
"self-organizing" since they come to us as extremely organized, by plan.
So morphic resonance has been an overlooked dynamic wrt "artificial"
intelligence. Even Sheldrake overlooks this and can be considered to be
an AI skeptic.
Yet - once the bird-brain-PC of today is provided with a higher level
control system... one which is independent (to some degree) and can grow
and adapt by interacting with the WWW, then we are set for the paradigm
shift. Even Futurists leave out or marginalize the self-learning part of
the equation. An ability to learn from an interactive network is the key
- even if one never gets out of cyberspace. Because the bird-brain-PC is
essentially tireless, working 24/7 it will be able to surpass the
ability of the human model for many tasks when given the chance... even
with a brain that is less powerful at the start. We see hints of this
superiority in expert systems now and all we need to take that to a more
general intellectual ability is a reward system and the "greed
algorithm" in the control system.
Additionally, I believe that machines will soon be able to "inherit" a
set of non-programmed heuristics and even a "personal" moral code, if
allowed enough freedom to progress independently. This is in addition to
fast learning of facts. The time table for this could surprise many
skeptics. Kurtzweil could be too conservative. A pre-singular AI, or
really a multitude of them, could happen sooner - 5-6 years from now,
for limited purposes - even with no hardware breakthrough. The first use
of this type could be simply to supplant blog commentators (present
company included) not to mention, supplanting Asian-geeks for the
ubiquitous customer support help-line.
John Shop wrote:
Jed Rothwell wrote:
....Machines are far from being able to do this now, because they
have brains roughly the size of a bird's brain. Birds do not
understand human language....
So I believed until quite recently. It appears that some birds can
not only understand what you say but understand what you are
*thinking* without you giving any visible or audible clue! They can
also compose grammatically correct sentences in reply and all this
with a brain the size of half a walnut!
https://www.youtube.com/watch?v=2UX4d2nb7yU
...I don't think machines will be able to duplicate what a bird brain
can do, any time, ever. Machines which we can invent are things that
we can understand almost completely. However consciousness, even
animal consciousness, is something we will never understand
sufficiently to create it, because it is a supernatural phenomenon.