Back in September 2025, I wrote a short post called Reality Is Contested.
At the time, I was trying to describe a shift that felt bigger than “AI content is getting better”. The internet was starting to feel unstable in a different way. Not just because things could be faked, but because people were starting to assume everything was.
A real photo looks fake. A human-written post reads like AI. A real video gets dismissed before anyone knows what it is. Actual expertise gets flattened into the same bucket as generated content.
Everything gets put on trial for being fake.
This short film captures that feeling well:
The unsettling part isn’t really the AI. It’s the conditioning.
We are getting trained to distrust our own perception.
Synthetic until proven real
For a long time, the internet ran on a rough assumption: if something reached you, it probably happened.
That was never fully true. Photos could be staged. Videos could be edited. People lied online. None of this is new.
What changed is the range. It’s not just images anymore. It’s voices, faces, writing, music, screenshots, calls, meetings, and context.
What feels different now is the default reaction.
Anything polished, emotional, impressive, strange, or too well-timed starts to trigger suspicion. Maybe it’s AI. Maybe it’s staged. Maybe it’s engagement bait. Maybe it’s a generated person, voice, story, or opinion.
That’s not paranoia. People are adapting to the internet around them. There is more synthetic content, it’s cheaper to make, and the incentives are obvious.
So people adapt too.
They stop trusting by default.
That’s the real shift. AI doesn’t need to fool everyone perfectly. It only needs to create enough doubt that people stop trusting the obvious.
The comments become the content
You can see the pattern everywhere now.
A video appears. The first argument isn’t about what happened. It’s about whether the thing is real.
Someone says it’s AI. Someone says it obviously isn’t. Someone demands proof. Someone else says it doesn’t matter.
The original thing gets pushed into the background. The argument over whether it is real becomes the content.
That feels like a preview of where a lot of the internet is heading. Less shared reality. More arguing over proof. Less “what do we think about this?” More “is this even real?”
That has a cost.
It makes people cynical. It rewards volume over trust. It makes good work feel suspicious because bad work can imitate the surface.
Human writing gets accused of being AI because it is too clean. AI writing gets praised as insightful because it copies the shape of insight. Real work and generated work start occupying the same feed.
That’s not just a content problem.
It’s a trust problem.
What still matters
I don’t think the answer is to reject AI.
The tools are useful. I use them every day. They help me write, code, review systems, test ideas, and move faster. I’m not interested in pretending otherwise.
The better question is not always “did AI write this?” or “did a human write this?” anymore.
That frame is too small.
Authorship is becoming more like a pipeline.
The human part is the taste, direction, constraints, lived context, and decision to publish. AI can help with the language. It doesn’t remove judgment or accountability.
As the internet gets more synthetic, the valuable things are the ones that are harder to fake over time.
Trusted people. Consistent identity. Visible process. Taste. Judgment. Reputation. Real work.
That last one matters most to me.
A generated architecture diagram is easy. A system that deploys, runs, breaks, gets fixed, and keeps evolving is different.
A generated opinion is easy. A point of view that stays consistent across months and years is different.
A generated profile is easy. A person with a trail of real work, real decisions, and real constraints is different.
The internet may get noisier, but continuity starts to matter more inside that noise.
Still contested
The irony is that AI may increase the value of being unmistakably human.
Not performatively human. Not “I added typos so you know I am real”. Actually human.
Specific. Situated. Opinionated. A bit uneven. Connected to real work. Connected to real context. Connected to a person who keeps showing up.
Reality is still contested. Maybe more than it was when I wrote the first post.
But the response is not to produce more content faster.
It’s to build more trust over time.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.