Are Algorithms the Magic Bullet?(calebelston.com)
calebelston.com
Are Algorithms the Magic Bullet?
http://calebelston.com/are-algorithms-the-magic-bullet-0
8 comments
In a word, maybe. I think algorithms is like the treasure map ... the way to get your insight , idea, solution done.
Everybody, everyday, everywhere have ideas, good and bad ones, brilliant and stupids ... but what makes the difference between what will works, make money and more important, get done, is the way you implement your idea. I have a personal notebook with dozens of insights but i have no idea how to implement most of then ... so what´s the point on having just the brilliant idea ?
But when you have an good idea and then write a good algorithm, you can get your idea very close from reality and that´s the most important thing.
So maybe, yes, the algorithms is kinda kings in a world where "get something useful done" is extremely urgent.
Good comment.
Saying that algorithms are the secret to the success of Google and the like is like saying that tunnelling machines are what made the London Underground popular - not quite true. Obviously, once they'd had the idea for the Underground, designing and building the machines for its construction became an absolute necessity which allowed them to create a far more wide-reaching network than previously thought possible - but this ignores the fact that people would still have used it if it had remained a hand-dug tunnel between King's Cross and Paddington.
In short, one could say that insight makes things popular while implementation makes them excellent - and that you should really be aiming for both.
Saying that algorithms are the secret to the success of Google and the like is like saying that tunnelling machines are what made the London Underground popular - not quite true. Obviously, once they'd had the idea for the Underground, designing and building the machines for its construction became an absolute necessity which allowed them to create a far more wide-reaching network than previously thought possible - but this ignores the fact that people would still have used it if it had remained a hand-dug tunnel between King's Cross and Paddington.
In short, one could say that insight makes things popular while implementation makes them excellent - and that you should really be aiming for both.
I'm not sure that's a good analogy. People would not use Google if it had "hand-dug" search results. It's that algorithmic edge that makes Google useful for finding things.
In a word, no. The netflix effort was many, many man-months of inspired work. And as Norvig points out in a recent paper, "simple models and a lot of data trump more elaborate models based on less data. ..."
Agreed - although Cuil's picture matching algorithm might have been a bullet to the head
i'm not 100% convinced this author has any idea what an 'algorithm' is.
Algorithms are simply one way of executing on insight, a way that is particularly well suited to large data sets, no doubt.
huh
Algorithms are simply one way of executing on insight, a way that is particularly well suited to large data sets, no doubt.
huh
Caleb (the author) here. That is a weak sentence for sure. My point here was that there are many sources of data, and not all are massive and easily understood by machines. But for those sources that are, algorithms provide the opportunity to make meaning out of the massive data.
I was on my way to post exactly that.