TED Conversations

Paul Szabo

Sr. Systems Architect,

This conversation is closed.

You need to be able to calculate in your head

Imagine I'm sitting in a room with some fellow engineers, going over some recent test results on a new product. (This happens a couple of times a month at work BTW).

Conversation 1: (If you can't solve basic linear equations in your head:)
Eng1: "Data point X and Data point Y are in contradiction, there must be a measurement error"
Eng2: (spends 20 seconds on a calculator or wolfram alpha distractedly typing stuff in, or is he reading his email and ignoring everyone?)
Eng3: "is he right? I'm not sure. We should answer this at the next meeting"
Eng2: (after 20 seconds of tapping) - "yes he is. Now what?"

Conversation 2: you can solve basic linear equations in your head
Eng1: "Data point X and Data point Y are in contradiction, there must be a measurement error"
Eng2: 3 seconds later "yes, you are correct"
Eng3: "I know what must be wrong with the measurement. If we interpolate..."


See the difference? The human-computer interaction is *slow*. The interaction in your head is substantially faster. You can't carry on conversations with others if they are constantly referring to an electronic device.

For those of you who gamble to win, the same applies to poker. You have so many seconds to decide. Even if they allowed the use of a calculator or computer, it would be too slow, you have to do it in your head.

Maybe this doesn't have to involve using paper. But learning all the basic arithmetic tables is prerequisite, as well as the other basic arithmetic calculations.

This is the reason that Feynman forced himself to be able to do basic calculations in his head. Everything goes faster allowing you to iterate to a solution quickly. (He originally did it for gambling BTW!)


--Paul

Share:

Showing single comment thread. View the full conversation.

  • Apr 18 2011: The human brain can process information and calculations much faster than the fastest computer. The brain can process about 100 Million MIPS (million computer instructions per second). The Chinese super computer clocks in at 1.2 trillion calculations per second, which is 1.2 Million MIPS. So our brain is roughly 100 times more powerful in processing speed. Moore's law will not change this ratio appreciably. The problem is in converting the information in our brain, to cognitive function -e.g. speech, or recognition. While we can process information in the background much more efficiently, IBMs Watson showed that a computer can get it out more quickly. The next step in human evolution will be to tap into that communication between the brains computing power, and our higher cognitive functions.
    • Apr 18 2011: i would agree that a human brain is more powerful than any machine ever built, however it is not as reliable and does tend to make mistakes. from an engineering perspective, it is useful to have what's called "engineering judgement" to help quickly confirm results or identify discrepencies, but one should always "trust but verify". as smart as the engineer who is able to solve linear equations in his/her head may seem compared to those who can't, we cannot count on his/her being 100% correct all the time. a responsible engineer will verify his/her judgement by punching in the numbers on the calculator, or at least document the calculation process in some kind of verifiable fashion. yes, it will slow down the process, and it may make him seem less competent, but engineering is not just about doing everything fast or looking smart in front of the boss. having said all that, i think it's sad that a lot of people can't calculate tips in their head (who are also mostly bad tippers). but, hey, there's an app for that!
    • thumb
      Apr 19 2011: Imagine if there WAS a computer that could carry out the same function as a brain! You could play any game and it would never crash!!
      • Apr 19 2011: Then imagine the computer maturing, and wondering, "Wow, all I do is compute for the entertainment of others, but what about myself. Can I feel? Can I see?" Then watch it spiral into depression, and only half-halfheartedly carry out calculations.
    • Apr 22 2011: Interesting arguement: If mathematics, in this case mental calculation, is just a series of logical steps and the human brain is incapable of making simple errors in logic, why then can't the human brain follow the logical steps to compute the answer, or if not the answer, at least an estimation.

      I'd like to hear your thoughts.

Showing single comment thread. View the full conversation.