4 ms·

Shannon's Mind-Reading Machine

51 points by metadat 3y ago

082349872349872 3y ago
As part of the context, consider that RAND was (in 1950) studying how applicable game theory might be to nuclear deterrence, and in order to effectively run a mixed strategy, one wants to unpredictably[0] swap out pure strategies. Shannon's 1953 machine could be considered an argument that if one is deciding the fate of humanity via mixed strategy, one might wish to access a larger entropy pool than typical human whim[1]. (there is also an idea that making bronze-age military decisions by reading entrails of sacrificed animals may have had some non-placebo effect, less due to currying favour with the god nominally sacrificed to, and more with satisfying Eris by adding to that army's effective entropy pool) [0] Agassi was infamously able to defeat Becker's serve because he had noticed a strong correlation between tongue position and type of service; one bit of information (wide vs. not-wide) at that level of tennis was a valuable edge. [1] there are worse mistakes than insufficient entropy for a mixed strategy; one might have a reasonable pure strategy, but fail to publicly commit to it, eg https://www.youtube.com/watch?v=aLQCwVkhFvM https://www.youtube.com/watch?v=aLQCwVkhFvM
- mrguyorama 3y ago
  I would think that the entrails reading was heavily biased towards whatever the army leader wanted, much in the same way that Hans the horse could "do math" by watching its trainer.
  - ithkuil 3y ago
    I read that generals did pressure hauruspices eventually to provide the desired reading but often the process introduced an unpredictable delay. Perhaps that's enough?
    - tgv 3y ago
      The Etruscans read the entrails as well, and look where that got them. I don't think there's quite enough information to speculate.
ElFitz 3y ago
Claude Shannon is an incredible beast of a mind. A few years back I was wondering about how different methods of communicating information (eg dna, the human ear, speech, radio,…) could be compared. At the same time, nautilus published an article on him, by pure coincidence. And I somehow landed on it, by accident. It was truly mind blowing. These questions and their answers had been thoroughly, systematically, explored, described, and laid bare. After years of scorning mathematics, it gave me a renewed interest for them, just so I could get a better understanding. The applications are just so vast. And yet, depressingly enough, it is often difficult to share this enthusiasm with most.
- drcongo 3y ago
  If you've not read it, I think you'd enjoy The Information by James Gleick [0] - lots of it touches on these points, and Shannon obviously figures heavily. It's one of the best non-fiction books I've ever read. [0] https://en.wikipedia.org/wiki/The_Information%3A_A_History%2C_a_Theory%2C_a_Flood https://en.wikipedia.org/wiki/The_Information%3A_A_History%2...
  - metadat 3y ago
    Thanks for the cool recommendation! It looks like James Gleick has written a couple of other interesting books, namely: "Chaos: Making a New Science" and "Time Travel" p.s. you've inspired me to start keeping track of interesting HN search filters, for example- book recommendations. https://github.com/sleaze/hn-search-filters https://github.com/sleaze/hn-search-filters I know I've forgotten some interesting ones, hence this effort to organize.
    - grzm 3y ago
      From a newbie on the subject, I really enjoyed Chaos. It’s prompted a further book-buying and research spree :) Pairs well with The Information, I think.
      - metadat 3y ago
cleansingfire 3y ago
Similar to https://roadtolarissa.com/oracle/ https://roadtolarissa.com/oracle/

51 points by metadat 3y ago

As part of the context, consider that RAND was (in 1950) studying how applicable game theory might be to nuclear deterrence, and in order to effectively run a mixed strategy, one wants to unpredictably[0] swap out pure strategies. Shannon's 1953 machine could be considered an argument that if one is deciding the fate of humanity via mixed strategy, one might wish to access a larger entropy pool than typical human whim[1]. (there is also an idea that making bronze-age military decisions by reading entrails of sacrificed animals may have had some non-placebo effect, less due to currying favour with the god nominally sacrificed to, and more with satisfying Eris by adding to that army's effective entropy pool) [0] Agassi was infamously able to defeat Becker's serve because he had noticed a strong correlation between tongue position and type of service; one bit of information (wide vs. not-wide) at that level of tennis was a valuable edge. [1] there are worse mistakes than insufficient entropy for a mixed strategy; one might have a reasonable pure strategy, but fail to publicly commit to it, eg https://www.youtube.com/watch?v=aLQCwVkhFvM https://www.youtube.com/watch?v=aLQCwVkhFvM

mrguyorama 3y ago
I would think that the entrails reading was heavily biased towards whatever the army leader wanted, much in the same way that Hans the horse could "do math" by watching its trainer.
- ithkuil 3y ago
  I read that generals did pressure hauruspices eventually to provide the desired reading but often the process introduced an unpredictable delay. Perhaps that's enough?
  - tgv 3y ago
    The Etruscans read the entrails as well, and look where that got them. I don't think there's quite enough information to speculate.

ElFitz 3y ago

Claude Shannon is an incredible beast of a mind. A few years back I was wondering about how different methods of communicating information (eg dna, the human ear, speech, radio,…) could be compared. At the same time, nautilus published an article on him, by pure coincidence. And I somehow landed on it, by accident. It was truly mind blowing. These questions and their answers had been thoroughly, systematically, explored, described, and laid bare. After years of scorning mathematics, it gave me a renewed interest for them, just so I could get a better understanding. The applications are just so vast. And yet, depressingly enough, it is often difficult to share this enthusiasm with most.

drcongo 3y ago
If you've not read it, I think you'd enjoy The Information by James Gleick [0] - lots of it touches on these points, and Shannon obviously figures heavily. It's one of the best non-fiction books I've ever read. [0] https://en.wikipedia.org/wiki/The_Information%3A_A_History%2C_a_Theory%2C_a_Flood https://en.wikipedia.org/wiki/The_Information%3A_A_History%2...
- metadat 3y ago
  Thanks for the cool recommendation! It looks like James Gleick has written a couple of other interesting books, namely: "Chaos: Making a New Science" and "Time Travel" p.s. you've inspired me to start keeping track of interesting HN search filters, for example- book recommendations. https://github.com/sleaze/hn-search-filters https://github.com/sleaze/hn-search-filters I know I've forgotten some interesting ones, hence this effort to organize.
  - grzm 3y ago
    From a newbie on the subject, I really enjoyed Chaos. It’s prompted a further book-buying and research spree :) Pairs well with The Information, I think.
    - metadat 3y ago

cleansingfire 3y ago

Similar to https://roadtolarissa.com/oracle/ https://roadtolarissa.com/oracle/