LinkedIn: Conclusion - Enshittification
Happy Independence Day, Friends!
Before I close the door on LinkedIn, I need to thank a couple of my friends for their patience, wisdom, and support. Genuinely stuck trying to name good things in LinkedIn, I used the "call a friend" lifeline.
- Aaron Fuller[1]
-
Aaron is a full-spectrum hacker. Others will spend weeks planning and arguing. Aaron quietly whips out his Jupyter Notebook — in any IDE. And has the answer in a day or two, while others still argue "worth it or not." Yet Aaron’s best magic comes from the fact that he’s never sidetracked by the noise of the world. And here was a bit of noise to deal with. I owe Aaron the big-picture view and the value mapping analysis.
- Michael Lipari
-
Michael is a people hacker — a human capital specialist across both recruiting and business development. On LinkedIn, he’s probably user 7 — the one who got many of us on. He used the platform from its earliest days: connecting people, referring people, brokering business development, hiring talent, and routing contacts through funnels and networks based on trust. I owe Michael the full value history of LinkedIn — from zero to now.
the TL;DR
The Australian Macquarie Dictionary[2] named "enshittification"[3][4][5][6][7]
its 2024 Word of the Year. The term describes the
gradual deterioration of a service or product brought about by a reduction in the quality of service provided -- especially of an online platform -- and as a consequence of profit-seeking.
It was selected by both the dictionary committee and the public vote.
"Enshittification" was coined by Cory Doctorow to describe how tech platforms evolve and <decay> as they begin prioritizing profit over user experience.
The process typically involves an initial phase of
offering good service to attract users,
followed by a decline in quality as the platform seeks to
maximize profits for shareholders.
not <decay>
Yet there is an important misleading point about decay in the formal description above by Cory. Consider a once-legendary insurance company in 1922, formed by a dynamic husband-wife duo. This company cut through market like Skjold-class corvette through Kattegat — with authority. Major investor acquisition later gives it nine lives and today it floats like FUBAR-class log in a canal.
Enshittification is not the insurance company’s rot. That rot was a product of time — a natural decay. This one is choreographed, planned, and intentional.
By <decay> Cory meant the user perspective alone.
the "Original Good"
Let’s return to LinkedIn (LI).
When it launched, it was disruptive across multiple fronts: a social-professional graph, a referral network, a reputation layer, and an early open platform.
Few remember this now, but LinkedIn once shipped true innovation — real value. Did you know LinkedIn once had a Connection Map — a full network visualization of your graph?
It was beautiful. But it was good for you!
- LinkedIn entered with a bang of value
-
-
2006: Public Profiles, Recommendations, People You May Know — not like today’s uncanny mess.
-
2007: LinkedIn Answers — early signs of real community building.
-
2008: Mobile version, Applications platform — a friendly nod to OSS.
-
- 2011–2014: Going Public and Feature Refinement
-
-
2013: Kills LinkedIn Answers, modifies 2005 feature Groups.
-
2014: Retires InMaps — a stunning network visualization tool.
-
It was still building for people, not metrics.
2016: Microsoft announces acquisition of LinkedIn for $26.2 billion.
— you remember that king whose touch turned everything to gold?
Yeah. That’s not what happened here.
the Hard Turn Port
-
2017: New desktop interface aligned with mobile — visual and interaction oddities not asked for by users.
-
2019: Launches Open for Business, enabling freelancers to showcase services.
-
2020—2022: Content Creation: Events, Creator Mode, analytics for newsletters and posts — new rating models.
-
2024—2025: AI Integration and Feature Saturation — rapidly versioned ranking and popularity models.
the "Tor Bazaar" and "Deep Waters"
In a way, it’s unfair to blame the acquisition alone — users were already being conditioned before the sale.
- Let’s just look at some of the most criticized removals
-
-
2013: Retirement of LinkedIn Answers — a Q&A system many users cherished.
-
2014: Discontinuation of InMaps — eliminated professional network visualizations.
-
2017: Removal of contact tagging and filtering — reduced user agency and precision.
-
2021: Termination of LinkedIn Stories — a short-lived engagement gimmick.
-
But the story is clearer when told from the perspective of the hacker collective.
There’s a constant tug-of-war between hackers and crackers.
Crackers are opportunistic criminals — they find vulnerabilities and steal data for resale. Hackers fix those vulnerabilities, closing doors before they’re breached again. These two groups despise each other and watch each other, and they are — like it or not — connected at the hip.
Tracking both sides tells you what marketing desperately tries to hide.
In the Tor bazaars where data leaks are traded by crackers, there’s a slower, colder undercurrent.
An undercurrent where both hackers and crackers scan the field — watching not just each other, but also corporate shifts and cultural signs.
Crackers look for things to steal, people to sell, and hackers to dox. Hackers look for danger signals — technical or ethical — that forecast breaches, surveillance, or rot.
So let’s walk the timeline of LinkedIn as the undercurrent saw it.
2011 — long Road to … Kakistocracy
In 2012, a devastating data breach leaked 6 million user passwords — slamming LinkedIn like a runaway train.
But the decay started earlier. By 2010, LinkedIn was a hacker powerhouse — few companies could match its invention.
The tech proves it.
Everyone knows Kafka by now. But that was just one spark from a bonfire.
The full constellation: Azkaban, Rest.LI, Pegasus, Helix, Samza, Gobblin — and that’s just what went viral. Dozens more surfaced briefly, shown in dim bars on chalkboards by shy, wild geniuses. Even today, LinkedIn still milks the output from that golden era.
My favorite was LiX — the LinkedIn Experimentation framework. I’ve yanked on every A/B and multivariate system under the sun — and I still found LiX inspiring. It was powerful and alive — a full hacker toolbag for curiosity and truth-seeking.
But brilliance has a price. That much magic paints a bullseye on your back.
And here’s the thing — what I just described is only possible with a perfect Community of Practice and a deeply bonded hacker collective. They were a tight-knit crew, not just coworkers.
And then came the investors.
In preparation for monetization, the company brought in a fresh crop of execs, PMs, and strategy heads to “set the course toward profit.” Which is polite-speak for: marginalize the hackers, replace weird with compliant.
Hackers don’t react well to that.
The first leaks to Tor were cultural, not about money. People cried, resigned, and whole teams walked. What was left was infrastructure too complex for the remaining teams to understand. IQ gaps you could sail a Samza through.
Upper management was unfazed. The mindset was: “Money solves all problems.” And if you’ve read the original Kafka — you know it’s a story of grief, alienation, and meaningless process.
Well — surprise. That’s exactly where LinkedIn Kafka came from.
So 2012 was inevitable.
And it raised no alarm bells, because Microsoft stepped in — with more money, and the cultural equivalent of zero-day exploits in human form. Microsoft didn’t invent the rot — they just accelerated it.
-
2017: 117 million credentials discovered for sale. Turns out, the leak had never stopped.
-
2018: First “bad API” scraping incident — technically legal, practically negligent.
-
2021: 700 million user profiles on sale — live and updated. Microsoft marked it “fixed”
-
2024: Scraping/Spoofing arms race — public evidence shows the leaks never ended.
Maybe now you understand Joe’s stance in our opening story.
He runs a security firm. And to hackers, incompetence is offensive.
By Joe’s moral code, LinkedIn is an open database. Because if they wanted it locked down, they’d have done it. He might even believe that leaking is the business model.
Contrary to popular belief, Microsoft didn’t break LinkedIn.
They just brought the tools — and the culture — to sweep the pieces under the rug.
Hey, everybody loves a good villain. And EvilCorp is still the best one in the game.
So now you have the backstory.
the "Reality on the Ground"
The important question is — what’s in it for us, the mere mortals?
Let us then start with the basic facts.
- LinkedIn is wildly profitable, operating on what I call the Loss Leader Model, currently
-
-
Premium Subscriptions (Gasoline): profitable with margin, expensing all infrastructure.
-
Marketing Solutions (Staples): $7B per year, in the current cycle.
-
Talent Solutions (The Candy): Microsoft will never tell you this.
-
Here’s what this means: LinkedIn runs like your neighborhood gas station. They might sell gasoline at a loss — yet allocate all operating costs to it. Then the staples, like milk and eggs, carry just enough margin to cover shelf space. But candy, cookies, and impulse buys are the cash cow.
In the LinkedIn world, even the gasoline turns a profit after paying for all infrastructure. Just think about that. No matter what they do to you or me — their net margins go up.
- We can immediately conclude that we have here
-
-
A monopoly on the professional identity market;
-
And, a hated incumbent.
-
Looks to be a very opportune situation for a hacker collective.
- However, mind you — we also have
-
-
A user and business data hoarder;
-
Pay-to-Play casino boss;
-
Bait-and-Switch gamer;
-
A Surveillance Paragon;
-
Business Development Operating System;
-
And a revenue Trojan Horse.
-
Sure, a few hackers could launch a Minimum Lovable Product in under a year — All the love LinkedIn once promised to be. And that would rocket to a hundred million dollars per year in a heartbeat.
But then a different game begins — one these hackers have never seen before. From there, a slow and hard rise to the top could change your company too.
- It’s worth remembering this
-
-
It doesn’t matter how good your software service is.
-
It only matters how good your business model is.
-
Certainly, given room, the same hackers could make LinkedIn great again. But there is absolutely no interest in doing so.
Case in point: I don’t use Windows, and so what?
the "Good in LinkedIn"
I was going to celebrate finding my long-lost friend on LinkedIn as my only positive.
She’s an incredible person and a great hacker I haven’t talked to — or coded with — for two decades. And I pinned her on LinkedIn — she responded. I was so happy I went and told my whole family about it.
But I didn’t find her on LinkedIn! LinkedIn did nothing for me here.
I found her in my LinkedIn data export — the ZIP archive you can request under privacy settings. Then I sliced through it using my DataSpell IDE and a few lines of Pandas. She had few connections, no activity, and no profile picture. In my natural interaction with the UX — she’d never show up. I’d have to manually filter the connections list, pushing buttons like it’s 2006. Well, I can do that much better in a Jupyter Notebook.
So that’s a win for me, not for LinkedIn. Yet it does bring us to a few legitimate use cases.
First, LinkedIn is a decent Rolodex if you’re disciplined. As my friend Michael pointed out — it’s the only place where you can reliably maintain professional contact with hundreds of people, especially if you’re constantly switching orgs, domains, or continents. If you’re methodical about curating your connections, the data export becomes your own CRM.
Second, the messaging system isn’t entirely useless. When someone changes their email, moves countries, or ditches their website — LinkedIn DMs are often the only channel still working. That assumes, of course, that the person logs in.
Third, it’s still a mildly credible business card. You can point someone to your LinkedIn when they ask what you do. Just don’t expect them to actually read anything.
But the real value — if we’re honest — is that LinkedIn is easily gamed.
Remember what Joe did in the opening story? He played the Sycophant Model. And he automated it. It was easy and wildly successful.
Perhaps the best feature of LinkedIn for a hacker is its exploitability.
You can bend the algorithm, train it, spoof it, monitor its pulse, and reroute its incentives.
It’s brittle and overfitted to synthetic behaviors. It’s like a corporate version of MySpace, with better CSS and worse taste.
And that — that — is the only solid option I see here: Game the game!
LinkedIn is an opportunistic business, not a cartoon villain — just like every healthy business should be. It functions by exploiting the psychological and social vulnerabilities of the American Employee Culture. And somehow, that exploit works just as well in many other countries. Founders and hackers aren’t their target. That’s why over 90% of my hacker friends don’t even have a LinkedIn account. (Except Joe. He now has a big one.)
But this leaves a large market open.
I’ve been running analytics since the experiment. And I estimate that the wide-open market — those permanently off the LinkedIn grid — is worth about 13% of LinkedIn, at a minimum. These are founders, indie hackers, competence shops — people who will never engage with LinkedIn. Right now, they have nothing of value designed for them.
- Here, it’s worth remembering how sales actually work in nature
-
-
There are people who will NEVER buy from you.
-
There are people who will buy with your effort.
-
And there are people who will ALWAYS buy from you.
-
That last group is your Natural Customers.
Every startup I’ve helped hackers raise — dozens, over the years — I hammer this point home:
-
Always start with your Natural Customer.
-
Get revenue early — as early as possible. Even before the product, if you can.
-
Expand outreach from there. Don’t stratify your market too soon.
-
And always, always seek out other hackers.
In this case, other hackers are the Natural Customer.
Let me know if you want to launch something. Maybe I’ll join you.
Happy long weekend!
Leave a comment