I Hate LinkedIn
I hate LinkedIn.
I want to say that clearly, at the beginning, before I spend the next several hundred words writing about it — which is, I am aware, a contradiction I will have to live with.
I hate the fake job postings. The listings that exist not because a company needs to hire someone, but because they want to appear to be growing, or they need to collect CVs for future reference, or some HR system requires a posting before an internal candidate can be confirmed. Malta is full of these. Most places are. You apply. You hear nothing. You were always going to hear nothing. The posting was never for you.
I hate the success stories. The ones that follow the exact same structure: "Three years ago I was sleeping on a friend's couch. Today I run a company with 47 employees. Here's what I learned." The couch is always mentioned. The 47 employees are always oddly specific. The lessons are always the same five things. The post always ends with "follow me for more." The engagement is always disproportionate to the insight.
I hate the fake humility. The posts that begin "I'm incredibly honored to announce..." followed by something the person is obviously not humble about at all, and shouldn't be — if you did something genuinely impressive, just say so. The performance of humility is more irritating than straightforward pride.
I hate the algorithm. Or rather, I hate the way people talk about the algorithm as though it were a sentient being with preferences and moods that must be appeased. "Post between 8 and 10 AM on Tuesday for maximum reach." "Use three to five hashtags but not more." "End with a question to drive comments." LinkedIn's own CEO has, on occasion, seemed uncertain about how the algorithm actually works. I have a theory that there may not be a coherent algorithm at all — just a system that was built, modified hundreds of times, and now behaves in ways nobody fully intended.
And yet.
I use it. I post on it. I check it. I measure impressions. I notice when something performs well and when it doesn't. I am, despite everything I just said, a LinkedIn user who takes LinkedIn seriously enough to care about the results.
Because there is no alternative. That is the trap. LinkedIn is the place where professional audiences exist, and if you want to reach professional audiences, you go where they are. The platform is terrible and it is also necessary and those two things are both true simultaneously and neither cancels the other out.
And Now There Is AI Slop
LinkedIn just added a "seems like AI slop" button.
In the first two weeks after launch, more than one million people clicked it. One million. In two weeks. On a platform where clicking anything requires a conscious decision to engage rather than scroll past. The volume of that response tells you something about how much accumulated frustration existed, waiting for exactly this release valve.
LinkedIn's chief product officer reported that members are experiencing 40% fewer views on content the platform has classified as AI slop compared to a few weeks prior. That sounds like progress. It is progress. It is also an acknowledgment that the problem was severe enough to require a dedicated button and a significant algorithmic intervention just to move the number by 40%.
What is AI slop, exactly? You know it when you see it. It's the post that begins with three rhetorical questions. It's the one where every paragraph is a single sentence, line-broken for dramatic effect. It's the one with the bullet points that restate the headline in five different ways. It's the one that sounds like it was written by someone who has read ten thousand LinkedIn posts and is averaging them. It probably was.
Pangram, an AI detection startup, found in a July analysis that LinkedIn was the most AI-saturated major platform — with more than 40% of its long-form posts flagged as completely AI-generated. LinkedIn accounted for nearly two-thirds of all content Pangram flagged as AI. Two-thirds.
This is not a small problem that LinkedIn is managing. This is a structural problem that LinkedIn created by building engagement mechanics that reward posting volume and consistency over quality and originality, and then watching as AI tools made volume and consistency trivially cheap to produce.
The Dead Internet Theory Is No Longer a Theory
The dead internet theory started as a fringe idea in tech circles: the notion that most of what appears to be human activity online is actually automated — bots, scrapers, engagement farms, AI-generated content — and that the genuine human presence on the internet is far smaller than the traffic numbers suggest.
It has graduated from theory to documented reality.
Cloudflare reported in April that for the first time, web traffic from AI surpassed traffic from human users. As of now, bots account for 61.9% of search requests. Humans account for 38.1%. A Pew Research Center study published last week found that of 10,000 web pages collected in July 2026, 10% showed significant signs of AI authorship — compared with about 2% five years ago. More than one-third of all web pages published after ChatGPT's release in late 2022 show evidence of AI authorship.
Spotify removed 75 million bulk uploads and duplicate songs over the last 12 months. 75 million. On a platform with an estimated 100 million total uploads. Substack launched an AI detection partnership with Pangram. Digg — the link-sharing site that was once a significant player in how the internet distributed news — shut down its app in March after being overwhelmed by bot spam within hours of its beta launch.
"Within hours, we got a taste of what we'd only heard rumors about," Digg's CEO wrote. "The internet is now populated, in meaningful part, by sophisticated AI agents and automated accounts. We knew bots were part of the landscape, but we didn't appreciate the scale, sophistication, or speed at which they'd find us."
The platforms that opened the door to AI-generated content are now trying to close it. The door does not close easily from the inside.
The Impression Obsession
Here is the thing about LinkedIn impressions that nobody wants to say out loud: a significant portion of them were always meaningless.
An impression is not a read. It is not a consideration. It is a moment when a piece of content appeared on a screen in front of a human being who may or may not have processed it before scrolling past. Before AI slop existed, the metrics were already inflated by the reality that most content on any social platform receives approximately zero genuine attention regardless of how many times it technically appeared in front of people.
AI slop didn't create the impression obsession. It exposed it. When 40% of long-form LinkedIn posts are AI-generated, the engagement numbers those posts accumulate don't tell you anything useful about whether a human being found the content valuable. They tell you that the algorithm rewarded the posting behavior that generated them.
The people who built personal brands on LinkedIn by posting consistently and engineering high-impression content are now discovering that the thing they optimized for was not quite the thing they thought they were building.
What LinkedIn Is Actually Trying to Do
LinkedIn is removing the "enhance your post" feature — the one that allowed users to rewrite their posts with AI — and replacing it with a proofreading tool. It is expanding profile and page verification. It is running the algorithmic intervention that produced the 40% reduction in slop views.
These are real changes. They are also reactive changes, made in response to a problem that LinkedIn's own design choices helped create. The engagement mechanics that reward posting volume, the lack of friction around publishing, the algorithmic amplification of content that generates comments and reactions — these design choices created the environment in which AI slop flourished. Removing one button and adding a flagging tool does not address the underlying incentive structure.
The deeper question is whether LinkedIn can maintain its value proposition — a professional network where real people share real perspectives — as AI makes it cheaper and easier to flood any platform with plausible-sounding content. The one million people who clicked the AI slop button in two weeks suggest the answer matters to its users. Whether the platform can engineer an answer before the problem becomes self-reinforcing is less clear.
The dead internet theory suggests that once a platform tips past a certain threshold of non-human content, human users begin to disengage, which reduces the human content, which makes the non-human content a larger proportion of what remains, which accelerates disengagement. Digg hit that tipping point within hours of its beta launch. LinkedIn is trying to ensure it doesn't.
I Still Hate LinkedIn
I will keep using it. I will keep posting on it. I will keep checking the impressions and noticing when something resonates and when it doesn't. I will keep measuring, because measuring is how you learn what works, and what works on LinkedIn — genuine perspective, specific detail, honest observation — is also what works everywhere else.
But I want to register, for the record, that the platform is deeply strange. That it is built on an incentive structure that rewards performance over substance. That its algorithm — whatever it actually is — amplifies things that generate engagement rather than things that deserve it. That the fake job postings are still there. That the success stories still follow the same template. That the couch is still being mentioned.
And that one million people clicking a "seems like AI slop" button in two weeks is not just a data point about LinkedIn's content quality problem. It is one million people saying, simultaneously, that they came to this platform to find something real and found something that wasn't, and they wanted someone to know.
That is, perhaps, the most human thing LinkedIn has produced in a while.
FreeMalta covers AI, digital platforms, and business ecosystems. For the full FreeMalta ecosystem, visit freemalta.com.
