The Rot at Facebook

Why you really should be reading Ed Zitron, and why you should break your Facebook and Instagram habits.

Not long after I sauntered over to Substack, I was introduced to a writer named Ed Zitron. He was/is a long time tech focused journalist with an original bias towards gaming (hence my first contact) and a healthy attitude towards the fuckery of the major platforms. His writing is now mostly on his newsletter “Where’s your Ed At?”, which used to be hosted on Substack, but in the aftermath of the Nazi situation at Substack, and the completely fucked up response by the founders, he moved it to Ghost. I highly recommend that you check it out.

One of his recurring themes is the “Rot Economy”, how the major tech platforms are really fucking everyone on the planet up. It dovetails nicely with the incomparable Cory Doctorow’s thesis on the Enshittification of the web; another theme that you ought to become acquainted with.

Ed’s latest post hit my inbox today, and it is an absolute BANGER of a post about the people who are destroying Facebook. What is shocking is how early in its history, long before Cambridge Analytica, before the decimation of news publishers (the shift to “Video”) and before the utterly shit-tastic failure that is the Metaverse. It started not long after the first opening of Facebook beyond college campuses, and the long and winding road to dominating the social media landscape.

I can’t do it justice, but I will pick out a few nuggets, and encourage you to go read this saga. Ed’s writing is not behind a paywall, but if like me you hang on to his every word, buy him a cup of coffee or two to show your appreciation and whack the subscribe button. I can assure you, he’s worth the space in your inbox.

First, here is the newsletter I will pull from for this: “The People Deliberately Killing Facebook”.

The Rot began early

Many people once they got on Facebook (FB) in the late aughts were delighted with the product. They were able to “connect” with friends and acquaintances, share stuff about their life, and keep in touch.

I remember this time, circa 2008 and 2009, and FB was becoming quite the time waster for me, but it was fun, and I got to keep up with “friends”.

But, even this early on, things were already getting rotten in Denmark…

Before the early public “fun” period when us non-students got access, there already were the seeds of the chaos being sown.

From the post:

We start our story in 2004, a few months after the company was founded, when Peter Thiel became Facebook’s first investor, putting $500,000 into the company in return for a 10.2% stake. As depicted in the 2010 Aaron Sorkin film The Social Network, Thiel was introduced to Zuckerberg by Sean Parker, the co-founder of Napster, his wounds fresh from a brutal beating by Lars Ulrich and Metallica. 

What was significant about this event wasn’t the people involved or the amount of money invested, but the terms of the deal — terms that would permanently and inextricably ruin Facebook. Parker, ever the advocate for founders, negotiated with Peter Thiel to allow Mark Zuckerberg to retain two of Facebook’s five board seats. When Parker resigned in 2005, he insisted that Zuckerberg be given his seat. 

Thus, almost at the beginning, Mark Zuckerberg was able to become the emperor of FB, ensuring that there was no practical way for him to be deposed or fired.

This is important, as it lays the foundation for the fuckery that would follow, over and over again.

In 2006, when revenues were about $6M per year, total users on the platform were about 9.5M, FB instituted the initial incarnation of the News Feed. Prior to this, it was much more of a discussion between friends, you would post on their walls, and they on yours. Sort of how the progenitor worked Myspace¹.

The new feature “News Feed” was not really a hit, and in fact was the start of the shitty-ness of Facebook.

In September 2006, Facebook would launch the News Feed — a relatively unsophisticated chronological feed of your friends’ status that is, on some level, one of the most important launches in modern software history, and a great example of a good product that got destroyed by executives that realized a focal point for their users is a great place to torture them with advertising. 

Don’t get me wrong. It had a rough start. The feed, which Facebook product manager Ruchi Sanghvi described as “quite unlike anything you can find on the web,” led to an immediate revolt within Facebook’s 9.5 million users, as it effectively showed a stream of literally every action you took on the platform, which users found “creepy” and “stalker-esque” according to a group made at the time called Students Against Facebook News Feed. 

Yeah, they captured everything you did, everything you clicked, everything you interacted with on the system was captured and shown to all your “friends”. Yikes.

Of course, they made this somewhat configurable, and hid a lot of it, but what they didn’t do, was to pare back the data that they gathered. This was really the first outside indication that they were hoovering all data, all the time.

This happened just before FB became publicly accessible (that is, when they removed the restriction that required a .edu email address to join). Yeah, the fuckery began before it really took off.

That same year, 2006, they turned down a $1B offer from Yahoo! to buy the platform, and they hired a former McKinsey analyst to be the “business” half, to Zuckerberg’s “techie” side. Sheryl Sandberg joined, and would be the adult in the room.

Yet, as the years went by, it became clear that she really wasn’t that great of a Yin balance to Zuck’s Yang, and instead she made it her mission to grow the platform. And grow it did.

In 2008, the platform had about 90 million users, and seemed to be stalled, when a VP of Platform & Monetization came up with a novel metric:

Chamath Palihapitiya, a former venture capitalist who was at the time Facebook’s VP of Platform & Monetization, came to Zuckerberg with an idea — that Facebook should focus on a new metric, monthly active users. This metric, which is now a cornerstone of most tech companies’ growth metrics, would be a measure of whether users were sticking on the platform, and boosting this number would mean that users were engaged with the product. 

In a meeting with Zuckerberg and Sandberg, Palihapitiya was asked to name it, and it came up with MAUs. Sandberg responded that they should “just call it growth.” 

At the next board meeting, Palihapitiya would present what Levy would call “aggressive growth techniques” that would double or triple Facebook’s user base, and then “use the platform itself as an engine for growth.”

And now, product managers like myself are expected to measure and drive growth in this “MAU” measure, the bane of my existence. But I digress…

He created a Growth team, who had almost unfettered responsibilities to drive that growth. The team:

Gleit, Olivan and Schultz are at the epicenter of almost every single choice that Meta has made to put growth above the user experience. Palihapitiya wanted their team — which also included early Facebook data scientist Danny Ferrante and Blake Ross, who previously co-created the Firefox Web Browser — to become “the Growth Circle,” what Levy refers to as “a power center in the company with special status and a distinctive subculture.” 

This led to the creation in 2008² of the “People You May Know” (PYMK) feature. I can attest that this even in the early days was eerie to behold. It surfaced people I literally hadn’t seen, talked to, or interacted with in more than a decade. And as the platform got larger with more people on it, the spookiness of this feature at times really freaked me the fuck out³.

From Ed:

He succeeded. In 2008, Facebook would launch a feature called “People You May Know,” a seemingly-innocent feature that would, as the name suggests, suggest people that you might know on Facebook. Yet this innocent seemed to be a little too good at its job, with reporter Kashmir Hill — who spent over a year investigating the feature for Gizmodo from 2016 to 2017 —  saying it “mined information users don’t have control over to make connections they may not want to make,” such as suggesting patients friend their psychiatrist, or outing a sex worker’s real identity to her clients.

Despite doggedly researching People You May Know, Hill never got Facebook to explain how it worked. Nevertheless, Steven Levy was able to get Palihapitiya to reveal one horrifying detail — that Facebook’s growth team would take out google ads on people’s names, targeting those who hadn’t joined Facebook with “dark profiles,” fake links that would suggest that somebody had already taken their name.  (emphasis mine)

This is some next-level fucked up shit, and not just spooky, but really fucking bad. Ed goes into some detail on the progression of the PYMK algos, with the tale of Lars Backstrom:

Backstrom’s LinkedIn notes that he “built PYMK (People You May Know)’s backend infrastructure and machine learning system” starting from September 2009 through February 2012. In 2013, Backstrom published a paper in tandem with computer scientist Jon Kleinberg called “Romantic Partnerships and the Dispersion of Social Ties: A Network Analysis of Relationship Status on Facebook.” The paper focuses on an algorithm that was able to independently identify someone’s spouse, according to the Times’ Steven Lohr, 60 percent of the time, and even able to calculate when a couple might break up. 

The paper hinged on the idea that mutual friends isn’t an indicator of a couple’s relationship status, but rather the “dispersion” of those mutual friends — the amount of mutual friends that are also mutual friends with each other. This is a seemingly sensible idea that, when framed as an algorithm made by the engineer who made Facebook’s extremely successful growth tool, feels far more creepy.

Backstrom, in a 2010 talk relayed by Steven Levy, said that People You May Know “accounted for a significant chunk of all friending on Facebook,” and that friends-of-friends are the most powerful part of the tool. People You May Know’s power wasn’t just that it found people that you knew well, but people you sort-of knew, offering you the tantalizing idea of getting close to them, and when you friended them, Facebook’s News Feed would make their content more prevalent in your feed, forcing an intimacy that may not have existed by making a new connection — no matter how tangential it may be — feel more immediate in your life. 

This subtle feature is responsible for much of Facebook’s growth, and was deliberately engineered by rot economists like Schultz, Gleit and Olivan to do so at any cost, even if it endangered the lives of children. 

It will not surprise you one bit that this feature became a fast favorite of the Pedo crowd, as more children flocked to the site (remember that age controls really didn’t exist in this era… and are still trivial to bypass).

According to the Wall Street Journal, in 2018 an engineer in charge of one of Facebook’s community integrity teams, David Erb, found that “the most frequent way adults found children to prey upon” was People You May Know. The Journal also reported that a few days later, Erb found that Meta was planning to add encryption to Facebook messages, something that would prevent the company from fixing the problem, threatening to resign in protest. He was placed on leave not long after, eventually leaving. 

Facebook introduced encryption to its messages, despite Erb’s stark warning that millions of pedophiles were targeting tens of millions of children. People You May Know was — and is — a dangerous tool, created, maintained and proliferated by people that now effectively run Meta. 

Because of course. And once they added encryption to Messenger, all that targeting and grooming became impossible to see and address.

Yay!

The Middle years

The growth curve had bent upwards, and as they approached their IPO (circa 2012), they were also approaching a billion users. Considering that the population of the Earth is about 7 billions, that is a breathtaking metric. This was before the dominance of the mobile platform. That means that 1 billion people were firing hup a browser on their computer and logging in to FB.

Staggering to think of it. Web-Scale was here. But how to monetize these people?

Back in 2008 and 2009 when I first joined, advertising was relatively primitive. You saw some ads in your news feed, but they weren’t terribly well targeted, nor were they very effective. Facebook wasn’t making bank from this, but it was enough of a revenue stream.

Until they instituted “Sponsored Stories”:

In January 2012, Facebook’s true rot would set in, with the launch of sponsored stories in Facebook’s News Feed, which Josh Constine, in a moment that should bring him deep shame, claimed he’d “rather see…inform [him] about the activity of friends than traditional ads that can be much less relevant,” a statement that frames exactly how little the tech media criticized this company. He also added one interesting tidbit — that Facebook had tested letting advertisers pay for sponsored content in the News Feed in 2006, but discontinued doing so in 2008, deciding that advertisers shouldn’t be able to show content in the News Feed unless it could appear there naturally, a statement that Constine wrote without a single hint of alarm.

In February, Lars Backstrom, the architect of People You May Know, would move over to manage News Feed ranking, and a few months later, Facebook would acquire Instagram for $1 billion dollars. In May, Facebook would have what would be considered a disastrous IPO, with Wall Street concerned about its lack of growth on mobile devices. After its first day of trading, the share price was barely above its debut, and by September, it had shed more than half its value. 

FB knew who you friended, knew what you viewed and engaged with, and armed with this information, they built a vast advertising empire. This was about when FB began tracking everything you did, all over the web. It is why my hosts file has about 1,400 lines in it to block the myriad tracking servers and services.

Fun fact: Google “tracking pixels”, an image that is a 1x1 pixel that many services and vendors on the internet embed into your messages and your browsing. You never see it, but with that, FB (and others) can follow you outside their ecosystem, gathering ever more data on you to use for the ad targeting. And since advertisers will pay bank to better segment their targets, this is why FB (or its new parent, Meta) is worth so much.

Like Zuck said, “I dunno why, people just give me all this data” from the earliest days when he was at Harvard.

Enough for Now.

This could get a LOT longer, but I will stop here. I do encourage you to read the entire post. There is info on the rise of click-bait, behavior farming, and manipulation of people to squeeze ever more value, drive engagement, and make more money.

I’ll leave you with this teaser:

This document — again, written in 2019 and published in 2020 — frames both how craven Facebook was about twisting its users’ habits, and how intimately aware it was of these problems. One commenter added, in a larger thought about addiction, that they “worried that driving sessions incentivized Facebook to make [its] products more addictive, without providing much more value.”

And that's exactly what Facebook has become: a gratuitous social experiment where the customer is manipulated to match metrics handed on from on high. 

Just fucking disgusting. I left shortly after the Cambridge Analytica fiasco, and I actively block all the tracking on my computers (it is not as simple to do on my phone/tablet tho).

Please check out Ed’s site, Where’s your Ed At? give him a subscription, and read him on this topic and on AI.

1. I will note that I never used Myspace, it was something that didn’t enter into my consciousness, but I knew it existed.

2. This is about when I signed up at the behest of my coworkers. Lord, what did I succumb to?

3. At one time, it surfaced a person to me that I had dated in the 1980’s before we lost touch, and we had no apparent common friends, but it made the connection by doing deem ML analysis of friends of friends.

---