6 ms·

"Knowledge dominates IQ" - along this lines, recently I have come to accept the idea "Speed dominates IQ": To explain it briefly, some one who can build fast a

"Knowledge dominates IQ" - along this lines, recently I have come to accept the idea "Speed dominates IQ": To explain it briefly, some one who can build fast and deploy can learn more as they come into contact with unknown-unknowns sooner and more often. Contrast this with someone with a huge IQ who thinks about things a ton, and builds really clever first solutions which breaks on first contact with reality/market.

tensor 13y ago
Running forward blindly does not help in research. During the talk, Alan even talks about this in the context of learning about the past to be able to come up with new ideas in the future.
- dredmorbius 13y ago
  True: failing fast only works well if you're failing at the right things, and learning from them.
- vidarh 13y ago
  On the other hand, there's a tendency with some people who "knows too much" to dismiss ideas that have existed for a while, even though it hasn't been widely disseminated. Sometimes running forwards blindly, while ignoring all the reasons not to do something is what finally gets innovations out to people.
  - bjelkeman-again 13y ago
    It get this at the moment. We have something new that the old hands say can't be done. We've proven it in the lab to our satisfaction and those that "know" just shake their heads. They do want more detail though. It is fun to take methods that are 80 years old and update them with new technology and have them outperform the state of the art.
    - PeterisP 13y ago
      Yup, in computing, there often is the situation where in 60ies or so they already figured out the proper way to solve some particular problem, but noted that is impractical due to the unimaginably huge resources required (such as gigabytes of storage or CPU comparable to a modern smartphone); while the 'state of art' of that problem is based on directions that were known to be strictly limited, but could be practically implemented given the hardware limitations of the time.
arethuza 13y ago
That reminds me of the idea that "The world is its own best model", credited to Rodney Brooks. Basically rather than having an intelligent agent rely upon a complex predictive model of its environment to explore the actions it can take a better/faster/simpler approach might be to actually interact with the environment and see what works - "intelligence without representation". http://en.wikipedia.org/wiki/Rodney_Brooks http://en.wikipedia.org/wiki/Rodney_Brooks

tensor 13y ago

Running forward blindly does not help in research. During the talk, Alan even talks about this in the context of learning about the past to be able to come up with new ideas in the future.

dredmorbius 13y ago
True: failing fast only works well if you're failing at the right things, and learning from them.
vidarh 13y ago
On the other hand, there's a tendency with some people who "knows too much" to dismiss ideas that have existed for a while, even though it hasn't been widely disseminated. Sometimes running forwards blindly, while ignoring all the reasons not to do something is what finally gets innovations out to people.
- bjelkeman-again 13y ago
  It get this at the moment. We have something new that the old hands say can't be done. We've proven it in the lab to our satisfaction and those that "know" just shake their heads. They do want more detail though. It is fun to take methods that are 80 years old and update them with new technology and have them outperform the state of the art.
  - PeterisP 13y ago
    Yup, in computing, there often is the situation where in 60ies or so they already figured out the proper way to solve some particular problem, but noted that is impractical due to the unimaginably huge resources required (such as gigabytes of storage or CPU comparable to a modern smartphone); while the 'state of art' of that problem is based on directions that were known to be strictly limited, but could be practically implemented given the hardware limitations of the time.

arethuza 13y ago

That reminds me of the idea that "The world is its own best model", credited to Rodney Brooks. Basically rather than having an intelligent agent rely upon a complex predictive model of its environment to explore the actions it can take a better/faster/simpler approach might be to actually interact with the environment and see what works - "intelligence without representation". http://en.wikipedia.org/wiki/Rodney_Brooks http://en.wikipedia.org/wiki/Rodney_Brooks