[extropy-chat] Self-Enhncmnt: data acquisition at high speed

[Acy James Stapp]

[ From: extropy-chat-bounces at lists.extropy.org 
[ [mailto:extropy-chat-bounces at lists.extropy.org] On Behalf Of 
[ Robert J. Bradbury
[ Sent: Wednesday, 21 January, 2004 19:57
[ To: ExI chat list
[ Subject: RE: [extropy-chat] Self-Enhncmnt: data acquisition 
[ at high speed
[ 
[ On Wed, 21 Jan 2004, Acy James Stapp wrote:
[ 
[ > I'm a fairly fast reader at ~1000+ wpm with fairly good
[ > comprehension on fiction and subjects I'm already acquanted
[ > with, and ~500 wpm on hard nonfiction. I've found that reading
[ > the same thing over and over until I'm quite bored with it
[ > produces massive gains in retention although it hardly affects
[ > comprehension.
[ 
[ But is there anything to significant to comprehend in fiction?
[ In contrast biology and history are largely retention while
[ inorganic chemistry and to a greater extent organic chemistry
[ require some retention and comprehension but to a large
[ degree practice of a set of rules.  Algebra and Calculus
[ are to a large extent the practice of a set of rules.
[ 
[ I'd like to see how different people would rate these
[ and any other subjects between retention/memorization,
[ comprehension and practice of methods.

Well, here are some of my recent readings and approximate speeds:

* "Engines of Creation", "Unbounding the future" - went really fast, 
  minimal rereading since I had already gathered much of the info 
  from other sources. Retention - easy, comprehension - easy,
  methods - NA

* "Nanosystems" - quite a bit slower, maybe 500-600 wpm and I 
  skimmed the math :) I loved the organization of this book 
  (Overview,  Explication, Conclusions) which makes it a good 
  learning experience for both top-down and bottom-up learners. 
  I'll be rereading this one in a few weeks. Retention - moderate
  (lots of new data), comprehension - good, methods - NA

* "Valence" (an old but fairly comprehensive treatment of
  molecular orbital and valence bond theory), "Advanced Organic
  Chemistry" - These are going quite slowly since my understanding 
  of differential equations is limited and I'm trying to understand 
  the math. Lots of rereading and my average speed is probably 
  as low as 300-400 wpm. Neither of these has exercises, which
  really help cement this kind of learning. Retention - poor,
  comprehension - moderate, methods - poor. It's going to take
  some study before I master quantum mechanics :) I will probably
  have to work on differential equations before I can truly master
  these.

* "Principles of Neural Science" - This is probably the hardest
  book I've read lately, due in part to it's poor organization
  and massive amount of information. This is as or more dense
  than either of the chemistry texts, but my limited biology
  background made it a lot slower reading. 200-300 wpm? Retention -
  poor, comprehension - good, methods - NA

* "Cambrian Intelligence", "How the Mind Works", "Words and Rules",
  "The Language Instinct", "Brainchildren", etc. I do a lot of
  reading on cognitive science, AI, and philosophy of mind so I
  go through these at nearly full speed. Retention - good,
  comprehension - good, methods - NA

[ 
[ > I'm sure that I could double my reading efficiency with some
[ > aids but simply rereading makes a tremendous difference.
[ > Taking notes would probably make a big difference as well,
[ > but I'm not quite that organized.
[ 
[ But as Harvey pointed out normal news content is data sparse
[ (I wonder if various news agencies realize this and whether
[ they are shifting summaries accordingly -- i.e. can I get
[ the news (or key points) from just the summary on Google News
[ or do I have to go to an article that may be 1 page on
[ some sources and 4 pages on other sources.  (Its kind
[ of amazing how many ways reporters/writers can repackage
[ the same press release.)

I deal with this by reading specialized news aggregators
like http://www.sciencedaily.com and 
http://www.newscientist.com/news/ first. Usually by the time
I see a story (abridged) in the popular geek press (Slashdot,
Kuro5hin, Fark) it's been several days since I saw it on one 
of the news sites.

[ This then becomes a question as to whether software tools
[ have enough knowledge/capabilities/rules to be able to
[ compress a story from a few dozen to a few hundred sources
[ into a dozen points about who/what/where/when, etc.?
[ Then if one reads/listens to the summary at 3x compression
[ one should be talking something like 10x or better than
[ normal methods.

I think this is a job for strong AI :) To really target
*you* this would have to know what you know, omit known
information from the news release, and gather required
background that you don't know, essentially understanding
it for you.

Acy