For ad targeting and recommendation algorithms, sure.<p>"Privacy," though isn't about Google shovelling reaction videos into my feed in order to keep me on the site.<p>It's more about every site requiring authentication and more information than they need about my identity, which in turn, gets sold to third-parties or gets stolen in a breach. And then can be used to impersonate me, craft scams targeting me, etc.<p>Agree though that recommendation algos are bad and over-used. They're designed to give the illusion that the system is curating content that you would like. An algorithm can't recommend something you would like. It hasn't sat down with you over countless dinners and nights out and got to know you; nor has it been off and read it's own books and seen movies on its own time. It's not going to drop off a worn copy of <i>Wide Sargasso Sea</i> because <i>that's what you need to read right now</i>. Recommendation algorithms are designed for retention and retention-only. They throw away almost all data about you except what you've clicked on recently.
> "Everyone loves collecting data, but nobody loves analyzing it later."<p>Analysing oceans of telemetry is not just <i>hard</i>, it's exponential-volume * permutating-international-linguistic-complexity hard.<p>The business model is therefore not to anything truly useful with useful data, the business model is <i>fake it hard by hoarding the data to make money</i>. If enough people believe that MAANG have real capability beyond non-semantic content correlation at scale, the money will (and does) accrue.<p>Say "A.I." breathlessly.<p>> chances are good that your pet ML application is an expensive replacement for a dumb heuristic<p>The probability approaches 1.<p>If the mere fact of holding data were put to a rigorous test, we would discover what everybody knows. [0]<p>> See, the problem is there's almost no way to know if you're right.<p>The dystopian truth is that it doesn't matter if the advertising empire is <i>right</i> or <i>good</i> or <i>helpful</i>. And the result is the grotesque coupling of what would otherwise be unequivocally brilliant technological advances: powerful mobile communication at scale and shared, connected, cultivated information.<p>- - -<p>[0] _ <a href="https://genius.com/Leonard-cohen-everybody-knows-lyrics" rel="nofollow noreferrer">https://genius.com/Leonard-cohen-everybody-knows-lyrics</a> _ "Everybody knows the scene is dead, but there's going to be a meter on your bed that will expose... what everybody knows." -- L. Cohen, S. Robinson
> Let's be clear: the best targeted ads I will ever see are the ones I get from a search engine when it serves an ad for exactly the thing I was searching for. Everybody wins: I find what I wanted, the vendor helps me buy their thing, and the search engine gets paid for connecting us. I don't know anybody who complains about this sort of ad. It's a good ad.<p>These ads are good in theory (and great when they work), but I find that, most of the time, they just result in me searching for one product and the first result being an ad for a competing product.<p>> There's another kind of ad that works well on me. I play video games sometimes, and I use Steam, and sometimes I browse through games on Steam and star the ones I'm considering buying. Later, when those games go on sale, Steam emails me to tell me they are on sale, and sometimes then I buy them. Again, everybody wins: I got a game I wanted (at a discount!), the game maker gets paid, and Steam gets paid for connecting us. And I can disable the emails if I want, but I don't want, because they are good ads.<p>This I totally agree with (though I might be biased, as a gamedev). These ads are awesome because I <i>told</i> Steam that I want these games and that I want ads about these games.
I live in a small country with multiple languages. Every site knows the country from the IP, and the language from my language headers. Even so, I don't think 10% of the ads I see give me something I can actually buy in my country, and in the correct language.<p>Youtube is the worst. I get the same 2 ads every few minutes, for weeks, generally for things not applicable to me.<p>The ad networks serve you whatever ads they happen to have laying around, and generally that's a 1 size fits all dumb generic thing. All those metrics exist not because they work, but so the marketeers can convince themselves and their bosses they are doing their job by shoving ads around. Nobody even has to look as long as the illusion someone is looking still holds.
It amazes me that these mega wealthy corporations don't understand uh... going abroad.<p>Like I know it's not so common on the US but here if you pop across to France for a week you get ads for everything in French. Despite my best* efforts I still speak abysmal French. You're not going to sell me shit in French (or indeed anything anyway).<p>I literally just crossed a border, I didn't become French, please.<p>This despite me being pretty lazy and generally letting my data get slurped left, right and center.<p>Perfect market efficiency strikes again...<p>*Citation needed
Something the author alluded to but didn't call out specifically is the "I'm already a customer" problem.<p>I can only really speak to Amazon and Facebook (via the web) because my uBlock configuration nips almost everything else, completely[0], and <i>most</i> advertising from those sites, as well.<p>But when an advertisement <i>does</i> slip through, it is <i>frequently</i> for a service I already subscribe to or a product I already purchased. In the latter case, I've had times where I've purchased something on Amazon and had the same product show up within a few minutes of the purchase as an advertisement on Facebook. So ... that was wasted marketing spend.<p>On the flip side of that, I'm <i>certain</i> Amazon knows who my girlfriend is and correlates our searches. There have been a handful of times when I was discussing gift ideas and she'll pull up a product and hand me her phone[1]. Sure, they pick it up when she sends me a link from the app or does some other action that can be correlated (i.e. we're both on the same WiFi, our phones are both reporting the same location[2]) but it's happened enough that I've started paying attention. On one occasion, she handed me her phone, we had no other digital back/forth and my phone was charging in the car (off, because I let the battery die). Later that evening, I'm on my laptop and <i>that product</i> was on the home page. There are other explanations, sure, but it'd be pretty easy to see that we <i>frequently</i> share one of two IP addresses when visiting their sites and start tracking those accounts as affiliated in some way for gift-giving purposes.<p>[0] I see product recommendations on Amazon and for some reason more than a few sponsored posts and the like on Facebook -- I don't use their mobile apps.<p>[1] I really, really hate using mobile apps.<p>[2] I don't actually know if the Amazon app tracks location ... I don't have it.<p>// FWIW: I really <i>did</i> laugh out loud at "Statistical Astrology" and I'm positive I'll be using that statement in the future.
Netflix just ended their 25 year long DVD by mail service. I subscribed to it a good chunk of that time (it was comparatively cheap), but didn't use it much over the last decade, because I'll admit I'm a dopamine fiend who likes that quick fix. But:<p>> Outliers are bad for business.<p>That's pretty much the takeaway here. Art (or anything really) that inspires, or is even more than about one sigma out, is not profitable. Sure, it's what you want, and you want to be erudite and trying to fill that final level on Maslow's hierarchy, but that doesn't guarantee quarter after quarter ROI for shareholders.
The author completely misses the second very important reason why someone might be against companies collecting sawths of personal data: government surveillance.
There are more and more cases where individuals get their phones searched without permission, their internet browsing history spied on, etc. Articles like this "I don't care" really don't help. In fact, they only do damage.
"The state of personalized recommendations is surprisingly terrible. At this point, the top recommendation is always a clickbait rage-creating article about movie stars or whatever Trump did or didn't do in the last 6 hours."<p>"There's no magic here. If you use ML to teach a computer how to sort through resumes, it will recommend you interview people with male, white-sounding names, because it turns out that's what your HR department already does. If you ask it what video a person like you wants to see next, it will recommend some political propaganda crap, because 50% of the time 90% of the people do watch that next, because they can't help themselves, and that's a pretty good success rate."
I'm glad this article mentioned Pandora - an oft-overlooked website that plays second or third fiddle to apple music and spotify, its recommendations have been surprisingly good and have been so for well over a decade now. It is one of the few entertainment subscription services I will always pay for unless some genius over there decides to ruin their algorithm with ML.
This talk by Tricia Wang along the same lines is worth a watch! The idea that we learn things by tracking or even with surveys leaves out the insights we get from actual human experience that's not purely quantifiable <a href="https://www.youtube.com/watch?v=WaOUJa9fjXU">https://www.youtube.com/watch?v=WaOUJa9fjXU</a>
> Netflix moved online, and the cost of a bad recommendation was much less<p>This was the most interesting point for me. I found it surprising that changing the cost of delivering a movie could so dramatically shift company-wide priorities.
I have no reason to believe that Netflix paid $1 million for a recommend algorithm and then threw it away.<p>Many DARPA projects from the 1990s were "cancelled" and then mysteriously showed up fifteen years later.<p>Either Netflix is lying about throwing it away, or we are supposed to believe a company that is fanatical about making money, to the degree of churning out 'sastificing' films and guiding users to them, is somehow also mr. butterfingers when it come to letting research grants slip through it's fingers and discarding the sucessful results..<p>Something smells fishy about this article. At what point do we admit tech companies don't operate for money and have their own private purposes?