Apologies for once again regressing into a favorite speculative subject
which has only passing relevance to alternative energy - the rapidly
approaching advent of AI (artificial intelligence).
Actually the relevance might become important, but anyway...
IMHO this eventual outcome of ongoing trends - progressing towards
machine-intelligence (artificial or not) will arrive 'sooner rather than
later' - and take many observers by surprise, because first of all,
computers are still at a rather low capability state, as of now ...
...and cannot even process normal speech very well - yet. Nor can they
process images very well - yet. etc etc. But the nature of progress in
all of these necessary sub-fields is proceeding apace with Moore's Law
advances in the underlying hardware, and - as is often the case with the
"tipping point" the end result will be spectacular, if only because the
audience for these stepwise advances has been lulled and accustomed to
the slow plodding pace - leading up to the tipping point. That is the
way 'critical mass' and 'emergence' works.
http://en.wikipedia.org/wiki/Emergence
http://en.wikipedia.org/wiki/Tipping_Point
... and also this is how Wiki became so smart, so quick ;-)
SIDE NOTE: The word "wiki" itself is an appropriate backronym and
metaphor for the rapidly advancing progress towards AI.
... since it is derived from the Hawaiian word meaning "quick" and
means: "What I Know Is". The wiki database servers are much larger than
one terabyte, however, but that is a function of the user base.
AND - have you (or any epistimologist) ever thought of Wiki in the
context of society's first (electronic) stab at a "group mind"?
Speaking of the terabyte, the reason that this broader implication of
(computer tech --> AI) came into focus this morning for me was that I
finally got my terabyte RAID disk array working, after figuring out that
the two new SATA hard disks, which were bought for a bargain, needed a
firmware upgrade. Sellers on eBay often do not mention why and how they
are able to sell items below retail, and this little 'problem' can be
one of the reasons... but still a minor inconvenience for getting 500
gigabytes of fast storage for less than a hundred dollars.
Thinking back to Marvin Minsky (now almost 80), founder of MIT's AI
laboratory; and author of most of the seminal thinking on AI and its
emergent philosophy - he had a one time opined that the human equivalent
of factual memory storage which would necessary for high level human
equivalence would be one terabyte. That was back when 100 megabytes was
a bargain at $1000. Today, for $200 (and a knowledge of DOS and the C:
prompt, if buying on eBay) anyone can get afford to have that one
terabyte on their desktop. The U.S. Library of Congress sez that as of
May 2007, the Library has collected more than 70 terabytes of data, but
if you were to compress this, weed out the BS and overlap, it would
probably be less than 10 Tbytes, indicating that everything worth
knowing is easily within the grasp of any AI.
If - that is - they could process it fast enough.
The necessary software, which is the "Son-of-XBox" (or PS3) processing
power (teraflop) will follow shortly... thanks to the demands of err...
gamers? Strange world, no?
In the meantime, any individiual can start to collect and store all of
the personally-important images, music, pdfs, and other data which they
identify with... That part is time-consuming. and the possibility of a
wet "brain dump" may not happen: at least in time for the next round, so
to speak ;-)
BTW I decided to name the disk array: "AB?" which is kind of a tentative
or quizzical play- on both the start of the "alphabet" (the original
word) and/or the techno-Karmic backronymic refinement of the "born
again" phenomenon ...
Jones
