Big-data daddy

Context: See Tru'eng anew, focus going forward, mathematics.

--

Big data? Yes, it's all one hears nowadays. If IBM has it in their ad stream, then you know it's about making money. What's next?

No one knows. But, it's nice having look backs that are insightful. Too bad that we don't have 20-20 hindsight (is foresight possible?).

An Atlantic article was a nice diversion from all those big-data dopes who push the smart (supposedly) devices that are really enmeshing agents. Dumb'd down, one might even claim (but, that's for another time). Hofstadter is from a former era. He's not the past, as his questions, and his work, are still relevant. The Atlantic had someone write about him who has a good feel for the issues.


So, reading this article is definitely worth the time.

   James Somer, November 2013,
                 The Man Who Would Teach Machines to Think

Atlantic, Nov 2013

Mentioning IBM was apropos: Deep Blue and Watson? Both of these got public attention, but they're dumber than a new-born baby, even one who might later be rated a less than average. Why say that? For several reasons. But, let's just say that both of these used brute force and the bullying of the data.

Aside: oh, don't some people act similarly?

You see, the idiots running after big data's worth (essentially, pushing toward the mean - with the large numbers involved) might appear to be obviously smart (yes, as in, the bigger the pockets, the smarter someone is - ignores ill-begotten gains, plus does not recognize the pervasive putrid-ness of near-zero realities). The techniques related to this type of work has been entrapping mankind in ways that do not portend well for the future. Of course, not all of mankind, as some have seemed to risen above the fray (which we could characterize several ways).

Like the article says, the approach of using large amounts of data might bring forth little nibbles of knowledge, but trying to get to truth? Won't work (we can, and will, describe how it ought to work).

---

These topics have all been discussed in this blog, at some time or another. Links may appear, at some point, here to those older posts. Not guaranteed, though. Perhaps, another approach is necessary.

In the meantime, it's nice to be reminded of Hofstadter's ilk (yes, those who are not chasing after that illusive buck (hugh accumulation) based truth set).

---

Same issue has another article of note.

       All Can Be Lost: 
            The Risk of Putting Our Knowledge in the Hands of Machines

It's nice that we can get assists from our tools. And, it our way to keep making better tools. Yet, to make the tool the focus is morally wrong (yes, golden sacks use the concept in terms of finance, so that gives me leeway); it doesn't take long, when looking around, to see the very deleterious results from the actions/decisions of the past couple of decades.

Aside: Hoftstadter's book, GEB, considers the importance of Godel's work. We've already addressed that, in part, under the guise of discussing computability, but many are not yet inclined to consider the ramifications (why? chasing after buck-based chimeras is too much fun).

Remarks:  Modified: 01/05/2015

11/11/2013 -- Yes, the Fed's data-driven illusion is on the table for discussion. 

12/31/2013 -- A popular post. 

03/02/2014 -- Analytics follow. Yet, we need qualitative analysis more (my put).

01/05/2015 -- Renewal, see Context line.