1. 7
  1.  

  2. 6

    This video contains an accurate (and entertaining) description of the fast Fourier transform. However, I’m disappointed by the exaggeration at the beginning of the video. To say that FFT is the “most important algorithm of all time” is somewhere between uninformed and grossly inaccurate. There are so many fields with at least as important of algorithms: cryptography (AES, RSA, Diffie-Hellman), sorting (quicksort, merge sort, heapsort), and search (PageRank), just to name a few.

    I’ve been watching Veritasium’s videos for about a decade now, and the amount of clickbait has gone up (which he briefly discussed) so much in the last few years that I don’t enjoy watching them anymore. I wonder if there’s anything platforms could do to disincentive clickbait like this (from an algorithmic point of view) while still surfacing popular content.

    1. 7

      maybe the ‘transformed’ is meant mainly as a pun.

      1. 3

        I’m with you, but I doubt there’s anything platforms can do on their own because they’d basically be fighting the profit motive.

        1. 3

          In a sense, you can imagine the world without each of these algorithms a bleaker place: without Diffie-Hellman, it’s very unlikely we could have online commerce or any digital privacy; without quicksort, a lot of other problems become really hard; without B-trees, mass storage would be really hard, and so forth. Without FFT, we probably would not have digital audio or video and it would be a major setback to science as well. Some of us lived through the time before PageRank, and it was absolutely awful. We’re pretty lucky to have the algorithms we do. I’m not sure I could pick a “most important.”

          1. 3

            To say that FFT is the “most important algorithm of all time” is somewhere between uninformed and grossly inaccurate.

            It’s a personal opinion but hardly out of the mainstream. From the FFT wiki page:

            In 1994, Gilbert Strang described the FFT as “the most important numerical algorithm of our lifetime” and it was included in Top 10 Algorithms of 20th Century by the IEEE magazine Computing in Science & Engineering.

            Regarding

            at least […] important […] algorithms: cryptography (AES, RSA, Diffie-Hellman), sorting (quicksort, merge sort, heapsort), and search (PageRank), just to name a few.

            Quicksort is mentioned in this list (pdf) (not sure if it’s the same as the IEEE one):

            • 1946: The Metropolis Algorithm
            • 1947: Simplex Method
            • 1950: Krylov Subspace Method
            • 1951: The Decompositional Approach to Matrix Computations
            • 1957: The Fortran Optimizing Compiler
            • 1959: QR Algorithm
            • 1962: Quicksort
            • 1965: Fast Fourier Transform
            • 1977: Integer Relation Detection
            • 1987: Fast Multipole Method

            Is PageRank published? Is is generally applicable in the same vein as FFT and quicksort?

            I wonder if there’s anything platforms could do to disincentive clickbait like this (from an algorithmic point of view) while still surfacing popular content.

            Probably not.

            1. 4

              Thanks for the reference to Gilbert Strang’s paper. I did not realize FFT was held in such a high regard by experts in the field. I still think that Veritasium’s statement “the most important algorithm of all time” is too imprecise to be reasonable. He left out two key points from Gil’s statement “the most important numerical algorithm of our lifetime.” From my reading of his paper, Gil is excluding all of the algorithms I mentioned because they don’t fall under numerical analysis. He is also excluding from consideration all of the numerical algorithms invented prior to the mid-20th century, such as Newton’s method or Gauss-Jordan elimination. The list of top 20 algorithms of the 20th century has a similar time-bounded restriction (and doesn’t try to claim FFT is #1).

              Is PageRank published?

              Yes, both as a publicly accessible technical report and in a journal paper. I agree that it is not as general as the other algorithms mentioned.

              1. 2

                I don’t disagree… but what do you expect from youtube edutainment? Sensational claims that can neither be proved nor refuted are bound to make people argue the point, and arguments drive traffic back to the video.