Monday, June 24, 2013

(This is the 14th post in the one-month challenge series)

I just wrote an entire post and deleted it before posting. It was about marks, privacy and parenting. I thought I had some points to say, but the more I wrote the more I realized this was leading nowhere. I was either roping in too many subjects or getting too vague about some. 

But that was not the real problem. I was myself not convinced with many of the things I had written. I believed in some of it, but my belief was vague. And once I realized that, I had to close my eyes and press delete. There is no point writing something about which you do not have much conviction. Which is what I call writing badly.

But, now you see, I don't have the energy to write one entire new post, so I will just say something real quick.. 

In my Psych class, we once talked about the "pigeon as a quality inspector". The idea was that pigeons are birds that can be trained VERY easily.  So you train a pigeon to compare a tablet (the one that has medicine, not the electronic one) with a prototype and reject all pieces that do not match. With some minimal reinforcements in between (a few grains and water every so many minutes), the pigeon can be used for this. The best part about this idea was that they showed that the pigeons were a lot more reliable (in terms of accuracy in discarding the bad products and letting in the good ones) than human beings.
Of course, they were not only efficient but cost-effective also. Just some food and water. No regular salary, no tea breaks, no 401k. You wouldn't find the pigeons chitchatting during work or engaging in politics and they can work even on national holidays. Apparently, the pigeons didn't get tired as quickly as humans either. Now, obviously no company would be ready to use pigeons as quality checking personnel. Even if there accuracy as higher than 99% and they needed less than a week to train. 

But wait. Isn't this kind of stuff automated now anyway? Of course it is. This paper is a veryold one. But the reason I mentioned it is because there are enough other manual jobs in the world which are no-brainers. And yet we could never automate them by training pigeons or mice.

Which is something for you to ponder. Aren't there some prejudices that prevail unequivocally against overwhelming evidence? If all the evidence in the world stll supported a mouse outperforming a human in a particular task, would you hire the mouse over the human?

Just something to think about..

Reference: http://psycnet.apa.org/journals/amp/21/2/109/ (Article not available for free.)

No comments:

Post a Comment

Comments are not moderated. However, spam (including irrelevant links to legal websites) will be deleted promptly.
Anonymous commenting is allowed, but it would be nice if you used SOME name just for ease of responding. Names don't link to your Google profile, so you won't be identifiable. Of course, it would be great if you used your true identity.
Also, as far as possible, please try to be respectful and criticize the idea, not the person (including other readers!)
Thanks!