Pluralistic: 06 Dec 2020


Today's links


Ad-tech as a bubble overdue for a bursting (permalink)

In my book "How to Destroy Surveillance Capitalism" I point out that the claims for Big Tech's behavior modification powers emanate from the companies' own self-serving boasts pitched to bring in new ad-tech customers.

https://onezero.medium.com/how-to-destroy-surveillance-capitalism-8135e6744d59

I point to the thinness of the external research on ad-tech's efficacy, and the replication failures of its research foundations on things like "sentiment analysis," "microexpressions" and "Big 5 Personalities" – the whole panoply of digital phrenology.

Meanwhile, there are undeniable, easily measured means by which Big Tech modifies our behavior that don't require us to treat marketing puffery as ground truth.

  • If you want to talk to your friends you have to use Facebook because Mark Zuckerberg is holding them hostage with monopoly tactics

  • Google uses monopoly rents to buy its way to search default on every platform, so the answer to every question you ask comes from Google

  • Apple gets to decide which apps you're allowed to use, who can fix your devices, and when you have to throw them away and buy new ones, thanks to DRM and lavish spending to kill dozens of Right to Repair initiatives

These are mass-scale, persistent behavior modifications that have nothing to do with psychological manipulation and everything to do with economic chicanery.

This week on the Freakonomics podcast, Stephen Dubner turns an economic lens on Big Tech's ad-tech boasts.

He talks to Steve Tadelis, an academic economist who once headed a program at Ebay to evaluate the efficacy of ad spending. First they tried eliminating "brand" advertising (that is, advertising buys for the word "ebay") and found that there was no drop in their revenues.

The logic behind Ebay buying ads for "ebay" is that if they don't, their competitors will, so a search for "ebay" will bring up links to Amazon. That happened…and people scrolled right past the Amazon ads to the "organic" Ebay link below the ads.

But Ebay also buys a bunch of keyword ads for products, like "guitar" or "washing machine" or "picture frame." They estimated 5% of their revenue came from these ads, and that every $1 they spent on them brought in $1.50.

Tadelis designed another experiment and found that these ads were actually responsible for 0.5% of their revenue – an order of magnitude less than their estimate – and that every $1 they spent generated $0.60 in losses. They cut $100m from their ad-spending.

But despite publication of these findings, the world increased its ad-tech spending. Tadelis attributes this to the fact that the major players in ad-tech are all incentivized to repeat the unsubstantiated tale of ad-tech's efficacy.

Ad-tech companies, publishers, and ad-tech buying consultancies are all compromised and unable to objectively assess whether ads work (cue Upton Sinclair: "It's difficult to get a man to understand something when his salary depends on his not understanding it").

And, of course, now Big Tech's critics have joined the ranks of those who insist that ad-tech works with spooky, devastating efficacy.

For evidence, all of them point to how much money the industry generates.

Why would people buy these products if they didn't work? Well, the obvious answer is, "That happens all the time." See:

  • Multivitamins

  • Hedge funds

But there's another answer: "it's a bubble."

A recent, excellent book on the ad-tech bubble is Tim Hwang's SUBPRIME ATTENTION CRISIS:

https://pluralistic.net/2020/10/05/florida-man/#wannamakers-ghost

Hwang tells Dubner that in a bubble, "the red lights are flashing but everybody in the industry just refuses to take a look at the real data."

Bubbles thrive on opacity and complexity. Think of the 2008 financial crisis, where the lack of transparency in "toxic assets" was compounded by their complexity, which led people – including "sophisticated" regulators and investors – to trust them.

They assumed that the manifest absurdity of the claims made by CDO salespeople must reflect their own ignorance. After all, all those OTHER people wouldn't spend trillions on derivatives if they weren't safe enough to buy yourself and exercise regulatory forbearance over.

But as Hwang points out, the ad-tech market is built on garbage. By Google's own reckoning, 60% of the ads that are charged for are never seen by any human being – literally the majority of the industry's product is a figment of feverish machine imaginations.

While Google's own research (and that of other Big Tech players) show that ad-tech works, independent researchers find the opposite: switching from "behavioral" (surveillance) ads to "contextual" ads only reduces clickthrough by 5%.

Behavioral ad clickthroughs are 0.01-0.03%, and much of that is bot activity.

The industry is opaque and incestuous. Ad agencies – nominally working for advertisers – get massive kickbacks from ad-tech platforms for bringing them business.

Proctor and Gamble – the company that invented the concept of brand ads – tried taking $200m out of its online ad spend and saw zero change in sales.

And yet, ad-tech spending continues to rise.

Hwang says we need a "punk rock" National Bureau of Economic Research, an org that will neutrally measure ad-tech performance and slowly deflate the bubble rather than bursting it, because an ad-tech collapse would kill ad-supported media.

All this is kind of a microcosm for the problems of economics in general. For decades, economics was dominated by the neoclassical idea of "homo economicus," a rational utility maximizer whose bad choices were good, actually.

(That's still the cartoon that undergrads get)

Advertising – especially brand ads – is grounded in the idea that irrationality is universal and exploitable, that you can trick people into paying a 50,000% markup by slapping a logo on a t-shirt.

But advertisers – and the industry – assume they are immune to irrationality, that they don't need to worry that they themselves will be suckered by slick sales-patter from ad tech, or their agencies.


The largest strike in human history (permalink)

Last week, the largest organized strike in human history shut down India. 250,000,000 people struck against Indian PM Narendra Modi's neoliberal reforms to the agricultural sector.

https://www.democracynow.org/2020/12/3/india_protests_modi_neoliberal_reforms

These reforms don't just remove the collective bargaining and price controls that protect the ag sector (which employs more than half the Indian working population), but also stripped multinational corporations and government of liability for harms to their workers.

All this while unemployment is at 27%, and 76% of rural Indians lack the funds to cover their basic nutritional needs. Meanwhile Indian billionaires have increased their wealth by 35% during the pandemic. India's richest man, Mukesh Ambani, has made $12m per hour since March.

The strike didn't just turn out unemployed people and farmers: the turnout was driven by acts of solidarity from ever sector of society.

In his discussion with Amy Goodman on Democracy Now, P Sainath offers some really important political context.

Modi has had a comfortable Congressional majority for two years and has three years left to go in his mandate, so why did he wait for the pandemic to make this far-reaching power-grab?

Sainath: "The reasoning was, these blokes are on their knees now. They can’t organize. They can’t hit back. And in fact, many leading neoliberal intellectuals, economists and journalists, editors, incited the government, saying, 'Never waste a good crisis.'"

It's shock doctrine shit, in other words. But Modi badly misjudged the moment: rather than being beating beyond resistance, Indians have been beaten to the sticking point, and will no longer be fooled by religious bigotry and neoliberal fairy tales.

Right wing movements around the world are grounded in the idea that some people are born to rule and the rest of us are born to be ruled over. Antimajoritarian philosophy isn't compatible with democracy, because it requires sustained turkeys-voting-for-Christmas to survive.

As India shows, the traditional tools of antimajoritarianism – xenophobia, sectarianism, armed violence – are unstable in the long run. Eventually there comes a point when you can't just shout "Muslims are scary!" at starving people and expect them to take that for an answer.

Indians have been slaughtered by both covid and mismanagement. They are at the breaking point. They are rising up.


Soviet computing graveyard (permalink)

In the 1970s, the Soviet Union started to clone DEC's PDP workhorse minicomputers, especially the PDP-8, which was replicated in the USSR as the Saratov-2. Today, the Saratov-2 is a distant memory, with not even a single high-quality photo of the system online.

Until now. Russian urban explorer Ralph Mirebs's photos of a "Soviet Computing Cemetery" (location undisclosed) that features the rotting remains of a Saratov-2 amid the ashes and fire-suppresant residue of a long-ago data-center blaze.

https://rusue.com/cemetery-of-soviet-computers/

The Saratov-2 was wild: it didn't have a microprocessor; rather, it was broken down into components, each in its own drawer: a 12-bit computing unit, I/O, RAM (ferromagnetic cubes).

Also present in the cemetery: an Electronics 100/25 – the Soviet version of the PDP-11 – and some DVK-2Ms (early personal computers).

The author recalls their own computer science education in 1993, when "one teaching DVK could distribute programs for a couple of dozen Spectrums through the network."

One of my last trips before the crisis hit was my visit to the Computer History Museum's boneyard – a massive warehouse filled with priceless paleocomputing remnants. Though the location is a secret, they let me take and post my photos:

https://www.flickr.com/search/?sort=date-taken-desc&safe;_search=1&tags;=computerhistorymuseum&user;_id=37996580417%40N01&view;_all=1

It was the end of an incredibly educational day I spent with Museum personnel, doing research for my case studies on the role that adversarial interoperability played in competition in the tech industry:

https://www.eff.org/deeplinks/2019/10/adversarial-interoperability

That day reminded me powerfully of my visit to St Petersburg's Popov Museum in 2006, back when my (now dead) great-uncle Boris Rachman was curator:

https://www.flickr.com/search/?user_id=37996580417%40N01&sort;=date-taken-desc&safe;_search=1&view;_all=1&tags;=popov

Soviet computing history is heroic in a way that's hard to put into words: the constraints of the era – political, economic, material – required so much ingenuity. Mirebs' photos for Russian Urban Exploration were the best thing I've seen all weekend.

(Image: Ralph Mirebs/Russian Urban Exploration)


This day in history (permalink)

#5yrsago Solo: Hope Larson’s webcomic of rock-n-roll, romance, and desperation https://memex.craphound.com/2015/12/07/solo-hope-larsons-webcomic-of-rock-n-roll-romance-and-desperation/

#5yrsago French Ministry of Interior wants to ban open wifi, Tor https://arstechnica.com/tech-policy/2015/12/france-looking-at-banning-tor-blocking-public-wi-fi/

#5yrsago READ: Kim Stanley Robinson’s first standalone story in 25 years! https://www.tor.com/2015/12/07/oral-argument-kim-stanley-robinson/

#1yrago A teenager describes his hilarious adventures installing a surplus, 1,500lb mainframe in his parents’ basement https://www.youtube.com/watch?v=45X4VP8CGtk

#1yrago The lawyer who caught UNC giving $2.5m to white nationalists orders the white nationalists to create a $2.5m fund for Black students or face a lawsuit https://indyweek.com/news/orange/t-greg-doucette-threatens-to-sue-sons-of-confederate-veterans/


Colophon (permalink)

Today's top sources: Naked Capitalism (https://www.nakedcapitalism.com/), Slashdot (https://slashdot.org/).

Currently writing: My next novel, "The Lost Cause," a post-GND novel about truth and reconciliation. Yesterday's progress: 515 words (86836 total).

Currently reading: The City We Became, NK Jemisin

Latest podcast: Someone Comes to Town, Someone Leaves Town (part 24) https://craphound.com/podcast/2020/11/23/someone-comes-to-town-someone-leaves-town-part-24/

Upcoming appearances:

Recent appearances:

Latest book:

This work licensed under a Creative Commons Attribution 4.0 license. That means you can use it any way you like, including commercially, provided that you attribute it to me, Cory Doctorow, and include a link to pluralistic.net.

https://creativecommons.org/licenses/by/4.0/

Quotations and images are not included in this license; they are included either under a limitation or exception to copyright, or on the basis of a separate license. Please exercise caution.

How to get Pluralistic:

Blog (no ads, tracking, or data-collection):

Pluralistic.net

Newsletter (no ads, tracking, or data-collection):

https://pluralistic.net/plura-list

Mastodon (no ads, tracking, or data-collection):

https://mamot.fr/web/accounts/303320

Twitter (mass-scale, unrestricted, third-party surveillance and advertising):

https://twitter.com/doctorow

Tumblr (mass-scale, unrestricted, third-party surveillance and advertising):

https://mostlysignssomeportents.tumblr.com/tagged/pluralistic

When life gives you SARS, you make sarsaparilla -Joey "Accordion Guy" DeVilla

1 Comment