In essence, the author seems to be applying a lot of the same principles used in machine learning to the way he thinks. The amazing part is that he's able to do this consciously.
I wonder if we'll ever be able to find a general solution for this problem, I've tried and ultimately failed in my last startup, but I think one of the biggest problem facing such startups isn't that this technology is hard, it's that people don't want it.
People enjoy processing more info than they can handle.
I've started down a similar path, but I'm much earlier in the process. Right now, I'm archiving and tagging as much as possible of what I read into Evernote. This helps me create a corpus of data to analyze later. Also, I'm trying to move away from using old-style RSS aggregators like Google Reader and toward a "daily briefing" consumption style, as in Paper.li and similar tools.
This article provides some useful ideas, though, several of which could migrate into my information processing workflow.
I think this path has flaws. First there is no idea why and how evolution developed consciousness inside our brains, and it's quite hard to simulate a black box only knowing the output.
Secondly, once this problem is solved teaching the machine information processing would be similar to teaching someone else which fails because it seems impossible to get all rationals (moving target) properly communicated.
Finally really _new_ news are rarely connected to existing data. To distinguish the interesting and surprising messages from just crap needs a different concept of how brains actually work.
3 comments
[ 5.8 ms ] story [ 18.8 ms ] threadI wonder if we'll ever be able to find a general solution for this problem, I've tried and ultimately failed in my last startup, but I think one of the biggest problem facing such startups isn't that this technology is hard, it's that people don't want it.
People enjoy processing more info than they can handle.
This article provides some useful ideas, though, several of which could migrate into my information processing workflow.
Secondly, once this problem is solved teaching the machine information processing would be similar to teaching someone else which fails because it seems impossible to get all rationals (moving target) properly communicated.
Finally really _new_ news are rarely connected to existing data. To distinguish the interesting and surprising messages from just crap needs a different concept of how brains actually work.