I’ve been on LinkedIn since its early days in 2003. Somewhere across of my 25 year career, it became a constant professional backdrop. I genuinely liked it. It felt useful in a way most social platforms never did.
I enjoyed seeing people do well. Promotions, career pivots, new companies, thoughtful essays about the future of work. There was real professional curiosity in the feed. You could follow interesting people and actually learn something. It felt signal-heavy and relatively low on spectacle. It didn’t feel like performance.
That’s why what LinkedIn has become matters to me. I’m not dunking on it. I used to value it. But it’s a clear example of what happens when engagement metrics and algorithmic incentives start shaping behaviour more than substance does.
The Metric Shift
Since Microsoft’s acquisition, something structural shifted. Growth became the dominant metric. Engagement became the currency. The feed shifted from “people I know” to “what performs.” When growth becomes the primary measure of success, optimization follows. Optimization rewards patterns. Patterns become playbooks. Playbooks become monoculture.
This is not another generic enshittification rant. Cory Doctorow’s framework helps explain the pattern, but what happened here was quieter. LinkedIn did not collapse. It simply followed the incentives of a public growth company.
The question is whether that optimization actually improved the product. Has the modern feed made professional exchange more thoughtful? More useful? More honest? It is hard to argue that it has. The platform works but it no longer feels aligned with the reasons many of us valued it in the first place.
The Repetition Problem
For me, the change shows up most clearly in repetition. I work in design and product, so my feed is saturated with AI discourse. Endless takes on how AI is replacing designers. Predictable headlines declaring that Figma is dead. Carousel posts built from identical templates. Different authors, same framing.
It probably performs extremely well. Shock headlines drive comments. Confident predictions drive reactions. Clean narrative arcs drive shares.
But after months of seeing the same topic recycled in slightly different forms, it starts to feel hollow. The issue isn’t that AI and design are unimportant. It’s that the format becomes more important than the thought. The monoculture becomes overwhelming.
When everyone optimizes for reach, originality collapses.
AI as Accelerant
AI didn’t create this dynamic, but it has accelerated it dramatically. When the cost of producing content approaches zero, volume explodes. When volume explodes, algorithms double down on what already performs. Convergence becomes inevitable.
AI makes it easier to mimic success. It makes it easier to generate a sharp headline or a polished take. But increasing volume does not automatically increase signal. More often, it compounds the noise.
The more people try to stand out, the more they end up sounding the same. AI isn’t the villain here. It’s an accelerant inside a system already optimized for engagement.
The Performance Economy
What we’re living inside isn’t just a LinkedIn issue. It’s what I would call the performance economy. In this environment, identity becomes content and sharing becomes strategy. You don’t just do the work, you narrate it for distribution.
When we optimize primarily for expansion, we should not be surprised when performance replaces authenticity.
Visibility starts to replace value as the thing being measured. This isn’t about blaming individuals. The incentives are structural. When engagement is the metric, performance is rational. People respond to what the system rewards. At scale, the system rewards frequency, spectacle, and replication.
And over time, that reshapes behaviour.
There’s also a personal dimension to this. Earlier in my career, this energy felt exciting. Building in public, sharing ideas, crafting a professional identity. There was fuel in that momentum. But as I’ve gotten older, with kids and a family and responsibilities that extend well beyond my career, the constant performance feels unsustainable. Real life introduces constraints: finite time, finite attention, finite energy.
The performance economy quietly assumes infinite output and constant presence. It rewards those who can produce continuously and publicly. Depth, quiet competence, and focused work don’t scale the same way. As you age, your tolerance for noise drops. Your career stops being your entire identity.
The Wrong Metrics
At its core, this is a metrics problem. When volume is valued over depth, visibility over value, and engagement over insight, the system bends toward expansion at any cost. LinkedIn is simply a clear illustration of what happens when the wrong metrics dominate a platform that once felt human.
And yes, I’m aware of the irony in posting this here. I’m inside the same system I’m critiquing. That tension is real. We’re all participating in structures shaped by incentives larger than any one of us.
Toward Principled Software
Growth and engagement are easy metrics. They are clean, quantifiable, and endlessly expandable. Trust, signal quality, and long-term health are harder to measure, but they may be far more meaningful. When we optimize primarily for expansion, we should not be surprised when performance replaces authenticity.
Nothing mysterious happened to LinkedIn. It became optimized. The more interesting question is whether we are willing to design and measure our systems differently.
I am not anti-technology, and I am not anti-AI. I use these tools. I build with them. But I increasingly believe we need to talk about principled software. Software that knows what it will not optimize for. Software that does not treat engagement as the ultimate measure of success. Software that respects the fact that human attention is finite.
I wrote more about this in Small Tech: The Need for Principle-Driven Software, where I argue that incentives shape outcomes more than intentions do. If we continue to design around growth alone, we will keep getting performance. If we design around signal, trust, and constraint, we might get something more sustainable.