The Internet we deserve

🌸 This post has bloomed and is unlikely to change.

I’m no AI doomer. I use ChatGPT to (hopefully) find answers when Google is failing me, to practice Italian or brainstorm weekend activities. I have a subscription to GitHub Copilot. I have built and will continue to build applications that use generative AI, both for myself and for my clients. It’s a tool in my developer tool belt.

Despite this, I have a growing feeling in my stomach - or maybe in my heart - that I can’t escape.

The Internet we know is dying.

Years ago I stumbled upon a conspiracy theory called “Dead Internet theory” that claims someone has taken over most of the internet with bots to manipulate public opinion and steer the zeitgeist. In this reality, most of the information you see online is generated with the sole purpose of affecting what you do and how you think. While this isn’t true, I fear that we are heading in a similar direction.

Every day I flinch more than before as I encounter obviously AI-generated content. It doesn’t matter if it’s on Bluesky, LinkedIn or a private Discord - just low-quality, soulless texts, images and videos without any substance. It’s as if the uncanny valley experience of reading the simulation in SubSimulatorGPT2 is slowly leaking into the real world, my world.

Unfortunately, there is no signs that it is slowing down. Services become cheaper to use and people keep chasing the promise of a return on their “investment”.

“Use a generated image and you will get more engagement!”

“Ask ChatGPT to write your LinkedIn content and you can produce 10x as much!”

“Automate your blog post creation with this amazing AI agent!”

That the output is mediocre at best doesn’t seem to matter. Mediocrity that floods our common digital space. Distinguishing what’s authentic will become harder and harder. As we adapt to this new reality both our trust and interest will drop, before we end up with an echo chamber of stupid robots talking to each other.

While authentic content drowning in an ocean of slop is a catastrophe in and of itself, it’s not the only problem. In order to satisfy the ever-hungry machines training new models, crawlers are relentlessly scouring the internet for new data to feed them. In our previous Internet we had functioning honour systems such as robots.txt that prevented unwanted bots from accessing our sites. Now a new tale of overly aggressive bots disrupting our most beloved sites, desperate conversations among sysadmins and massively increased side-project costs emerge every other day. This isn’t The War Against the Machines that the Terminator movies promised us, but it’s the war we got and I am afraid that we are losing.

In the midst of all this, the saddest realization is that these are decisions made with intent. Someone decided that ignoring robots.txt was worth it. Someone decided that letting crawlers pummel every site was the right move. That copyright doesn’t matter. That the environment has to be neglected in the name of imaginary progress. These are decisions made by people. There’s no Artificial General Intelligence to blame.

This is not the Internet that we want,

but maybe,

it is the Internet that we deserve.

Last update: April 08, 2025