I visited my Amazon app last week and was amazed to find the entire scroll packed full of thoughtful recommendations for what women’s apparel I should consider purchasing that day. While I entirely understand that this may have its appeal—and totally support whatever self-identification individuals make in their lives—the simple truth for this aging boomer (me) is that (1) I am (as the saying goes) heteronormative with (2) no fetishes that I have detected to date. But there were all these clothes on my Amazon app and they were 100% women’s items! I’ve been shopping with Amazon since 1997, have never purchased a single item of this type (nope, never secretly wanted to either), and you’d think that with nearly two decades of consumer purchasing data from me directly—information I have given them because I like using them for books, CDs, computer equipment, some bulk or hard to acquire foods, the occasional pair of men’s pants or a definitely male UA workout shirt—I could be spared this bombardment of off-kilter suggestions.
But there they were. And there I was, wondering why I had been provided this menu of stuff I was not going to purchase—ever—and definitely not wear—ever—and that my cat (the wonderfully talented eating, sleeping, and pooping home entertainment center known to me as Emma and known to herself as some derivative of “meow,” I assume) was too small to enjoy, although she is a female and would probably enjoy sleeping on them and eventually rip them to shreds with her inadvertent claw catches (how, by the way, since I’m here, is it that cats are so agile and intelligent in so many ways but can’t seem to figure out how to unhook their claw(s) from my shirt or pants fabric or the chair cushion, etc.?).
It is equally mystifying on Netflix. I log in and there are the films and TV shows they recommend, most of which I wouldn’t watch if they paid me (full disclosure: they don’t; I pay them) and would not recommend to my least intelligent acquaintance (or our state’s governor—same difference).
“May we remind you, kind customer, that our completely useless comedy series starring the nearly always awful Adam Sandler is available for your viewing pleasure?” I suppose you may but I sure wish you knew me better through my long history of NOT choosing Adam Sandler in anything other than Punch Drunk Love as I think he is an unfunny pillock of the worst kind (has anyone else in the U.S. noticed that the British are WAY more inventive with their insults than we are? Their lists just go on and on and we should purloin them to our version of the language as quickly as possible! Note to the wary: some of them already have alternative meanings in “American” and should not be used here or may result in a kick to the yarbles (not British slang but a word created by Anthony Burgess, so kind of British anyway).
To be clear, my film tastes tend to go towards serious drama topics, including well-done period pieces, dramas about demographics I know little about (films from other countries and social strata, here or elsewhere, etc.), really dark British detective series (Happy Valley, Luther, Line of Duty, in which almost all of the characters are having troubles at work and home), in other words, stuff I can think about, mull over, learn stuff from in one way or another. These are NOT areas that are best summarized by the two nouns Adam and Sandler. I also like some comedy (the sillier the better (e.g. W1A, Red Dwarf, Monty Python), some stand-up (e.g. Chris Rock, Dave Chappelle, Ali Wong, Iliza Schlesinger, Louis C.K.) some others that don’t spring to mind (all of these get down and dirty, btw)).
This kind of thing happens with social media platforms as well. Many of us are dutifully entering our personal likes and dislikes into these things. Our information is harvested, transmogrified into values of some type, sorted into the demographic to which we unwittingly belong, and ads are summoned up that are supposedly tailor-made for our eyes only. To misquote both Robert Oppenheimer and the Bhagavad Gita, “Now we have become data, the destroyer of worlds” (Q1: does one place misquotes in quotations? and Q2: is it wrong to take such a serious quote and make it about “Big Data?”).
It would be one thing I suppose if our data doppelgänger would provide endlessly useful, on-point suggestions. It is another thing altogether when our data are so incredibly misinterpreted as in the couple of examples I’ve provided above (the link on data doppelgänger is a legitimately interesting article on the topic I am whining about today; please read).
The behavior of search engine algorithms is at least as odd as the results described heretofore. I search for appropriately odd images for my posts and select “labeled for reuse” through Google as many of you do. I searched for an image for “A Cold House” recently and was immediately presented with the following item (I’ve given her a little cover as she was a bit too revealing for my imagined readers):
Why would this be an image suggested by the search phrase “A Cold House?” There were many similar images provided as suggestions that day but even on days where the thong-enhanced buttocks of a sailor are not among the suggestions, there are many suggestions that make absolutely no sense at all! These Google suggestions are not in the same realm as those provided by Amazon or Netflix but there should be SOME correlation between the search string and the results, shouldn’t there?
(To be a tiny bit fair, Google seems to have refined their algorithm since my initial search and although this young lady is still offered up as “a cold house” for some reason, many other scantily clad women who initially appeared have made their way elsewhere.)
I am puzzled every time I do such a search and am presented with random stuff that does not meet my needs. This time, the prompt gave me an opportunity to vent a little. It’s a little rant-y and I have no useful suggestions, except that jobs for data-mining large data sets, i.e. jobs focused on “big data” seem to be on the rise and this suggests that developing skills in whatever that all is might be useful… until they aren’t.
Given that virtually any article you read about Amazon, Netflix, or Google touts their ultra-refined customer and/or search algorithms, you would think that better results would be forthcoming.
That has not been the case for me.
Featured image (to be fair once more time, this illlustration is about a computer science algorithm problem called the dining philospher’s problem that may or may not have anything to do with consumer algorithms).