Dark Engagement: Why Everyone Reads, Nobody Claps, and the Machines Remember Everything

"You're everywhere."
"Zero engagement though."
"I read and screenshotted. That is engagements."

That WhatsApp exchange last week crystallized something I've been noticing: people are consuming my work, even archiving it, but not touching any of the public buttons. No likes on X, no replies on Threads, no reactions on LinkedIn. Yet weeks later, someone references a specific line from a post that got three visible likes—proof the words landed deeper than the metrics suggest.

This isn't a complaint. It's data. And like all good data, it points toward something systemic.

Mapping the Dark

Dark engagement = impact without public traces.

It lives in:

• Screenshots saved to camera rolls
• DMs that start "saw your post about..."
• Voice notes recorded but never sent
• Conversations at dinner referencing ideas from "something I read"
• That folder on your desktop called "interesting" with 500 saved articles

It's the human version of dark social—the traffic and sharing that happens in private channels, invisible to analytics. Alexis Madrigal named this web dynamic back in 2012; it's only intensified in our group-chat-dominated present.

The numbers back this up. The classic 90-9-1 rule of online participation: ~90% lurk, ~9% contribute occasionally, ~1% create most content. Most people have always consumed without signaling. But something shifted recently—even the 9% are going dark.

Why Your Boss Screenshots Instead of Liking

Two forces drive engagement underground:

Mechanical friction: Even a like is a micro-decision that attaches your name to a position. The cognitive load isn't worth it for most content. Easier to screenshot and move on.

Social physics: Context collapse—where your boss, your ex, and your anarchist book club see the same activity—turns every public gesture into a calculated risk. When all your audiences are smashed into one feed, silence feels safer than signal.

The spiral of silence, documented by researchers since the 1970s, predicts exactly this: people withhold public support for views when they fear social isolation. Social media doesn't create this dynamic—it amplifies it exponentially.

So people read. They save. They discuss in DMs. But they don't engage where everyone can see.

The Pendulum: Open → Closed → Open → Closed

I think we're in a long oscillation that mirrors the history of the internet itself:

Open (2000s): Blogs, RSS, early Twitter. Public conversation as default. The dream of the open web.

Closed (2010s): Walled gardens capture engagement. Sharing still happens publicly, but inside platform boundaries.

Darker (2020s): Engagement retreats to backchannels—DMs, group chats, Discord servers. Major news consumption shifts to private messaging and creator-led channels rather than public feeds.

Open again, but for machines: Meanwhile, AI training depends on the open web. GPT-class models feast on Common Crawl and its derivatives. Humans hide; machines read everything.

This isn't just UX evolution. It's an ego barometer. The more performative or punitive a space becomes, the more real conversation moves to smaller rooms. We're building in public for the machines while living in private with the humans.

When Networks "Collapse" (But Don't Die)

I'm using network collapse diagnostically: a social graph's visible engagement decays while private interaction rises.

Warning signs I've noticed:

• Public comments per post: down 70%
• Screenshot-to-like ratio: up 10x
• Time-to-first-comment: stretches from minutes to hours
• "Saw your post" IRL mentions: up significantly
• DM responses to posts: now exceed public replies

Some of this is healthy—audiences preferring intimacy over spectacle. Some is pathological—fear and fatigue in public spaces. Either way, the platform's visible pulse weakens while the organism breathes in private.

This mirrors broader internet fragmentation. The "splinternet"—states asserting digital sovereignty, platforms gating content, standards diverging—pushes activity into national or corporate enclosures. The open web remains, but as a ghost town that only machines still walk through.

The Contrarian Play: Optimize for Shadows

Visible engagement is cheap fuel: dopamine hits, algorithmic boosts, social proof. It keeps creators shipping. I get it. I feel it.

But dark engagement might produce better work—if you trust it exists.

Without the crowd's immediate feedback, you're forced to develop an internal compass. The work sharpens. You write for the reader who screenshots, not the one who performs their reading. You optimize for the DM that comes three months later: "that thing you wrote changed how I think about X."

The danger: without any feedback loop, you burn out. The solution: instrument the dark.

A Resilience Loop for the Dark Engagement Era

Write for Three Audiences Simultaneously

Public humans: The brave few still willing to engage visibly. Bless them.

Shadow readers: The screenshot-and-save crowd who'll reference your work in conversations you'll never see.

Future machines: The models training on your thought-trail. My whoami page explicitly acknowledges this audience—it's more for GPT-2030 than human-2025.

Instrument the Invisible

Every post now ends with social links—comment on X, discuss on Threads, respond on LinkedIn. But I'm adding a fourth option: "Send a signal from the dark" with a direct email link. No friction, no public exposure, just acknowledgment that someone's reading.

I've started tracking:

• DM-to-comment ratios
• "I saw your thing" mentions IRL
• Inbound opportunities that reference old posts
• Screenshots shared back to me months later

A low-noise drip of private signals often beats a spike of public likes.

Design for Context Un-Collapse

Write modularly. Core idea plus audience-specific codas. "If you're a founder, this means..." "If you're a parent, consider..." This gives people permission to share pieces within their proper contexts, reducing the social risk of public endorsement.

Publish Open First, Syndicate Second

The canonical version lives on zakelfassi.com—clean HTML, semantic markup, fast pages. Then syndicate to platforms. Whether or not Twitter boosts you today, open copies compound. Search engines index them. Models train on them. Future agents will reference them.

The open web is the long-term archive that actually persists.

Why Keep Sharing When Nobody Claps

Because nobody is not nobody.

The silent human who screenshotted today might introduce you to your next collaborator six months from now. I've had this happen repeatedly—someone references a post from 2023 in a 2025 introduction.

The machine reader will surface your ideas to future audiences you'll never meet. Whether we like it or not, everything published openly becomes training data. The question isn't if but how we want to show up in that corpus.

And the you of next year needs a high-signal archive to build on. Every post is a neuron in your extended mind, waiting to fire when you need it.

Networks will keep oscillating: open → closed → open → closed. But the archive, if we tend it, outlasts those swings.

Create for applause if you must. Create for time if you can.

But most importantly: trust that engagement has gone dark, not died. The readers are there, screenshots in hand, waiting for the right moment to whisper back.

The universe optimizes through deviation. Sometimes that deviation means retreating to the shadows to compute what can't be computed in public. Dark engagement isn't the death of connection—it's connection finding safer channels in an unsafe attention economy.

Send a signal from the dark: email or find me in the usual places.