“Watch out, you might get what you're after...
I'm an ordinary guy, burning down the house.”
— Talking Heads
The 1984 Jonathan Demme masterpiece Stop Making Sense of Talking Heads performing Burning Down the House live is better but I digress.
The real issue is the fact Google's burning down it's house in order to feed it's LLM and stiff creators and people needing information unfiltered and not processed into AI slop.
To whit.
The boffins in Silicon Valley finally got exactly what they wanted: total control of the data. And in the process, they are currently burning their own house to the ground.
For twenty years, the internet operated on a basic exchange. Human creators (publishers, writers, artists) provided (the content). Google provided (the traffic). It was a symbiotic relationship because if vendors made money from people doing the search, Google would get a cut of the sale. Vendors were happy because there was no friction.
I'm looking for a hat. A Derby or Bowler. I type in the prompt. Where can I find a Derby or Bowler hat? The response, XYZ companies showing product, pricing and availability. I select the company and buy the hat. Google gets a cut and the company in question is happy as Google's cut is akin to CC companies getting a cut.
You can still do this but it's getting harder as Googles push to feed its LLM and not us shows, in realtime, that the system's beginning to break down.
Now the AI Overviews and the LLMs take presidence. Google's gotten greedy. They decided they didn't want to route traffic anymore; they just wanted to scrape the content, summarize it with an AI, and keep the user on their own site. They achieved a "Zero-Click" search rate of over 70%.
They found the goose that laid the golden eggs, and in their infinite, ego-driven wisdom,
they slaughtered it to get all the gold at once.
they slaughtered it to get all the gold at once.
Now, they're starving.
1. Model Autophagy (Eating Their Own Exhaust)
Because Google and the AI labs choked off the traffic, the publishers are dying. NPR called it an "extinction-level event," and massive chains are threatening to block Google's crawlers entirely.
With human creators abandoning Google's ecosystem, the AI labs have hit a catastrophic data wall. They are running out of fresh, high-quality human intuition to train their next generation of models. Their solution? Train the new AI on the output of the old AI.
In computer science, this is triggering Model Autophagy Disorder (MAD). When a pattern-matching machine trains on synthetic "slop" instead of organic human truth, the data degrades. It loses the nuance, the lateral leaps, and the human spark. It homogenizes into absolute nonsense. The machine is literally choking on its own exhaust.
2. The Internal Mutiny (The Gulag)
But it’s not just the data that's breaking down; employees inside these companies are shattering.
The engineers now realize they have no purpose. Inside Meta, thousands of employees have been forcibly reassigned to perform mindless human-in-the-loop AI training tasks. Internal leaks reveal engineers referring to their own applied AI units as "The Gulag," stating: "I have no purpose in life. It feels like I'm just given these random tasks. I don't have agency anymore."
Over at Google DeepMind, over 600 employees recently staged a fierce internal revolt over executives selling out their safety protocols for military contracts.
The people who built the machine are realizing they are trapped inside it.
And lastly ...
Channeling Orwell, USA Today is losing customers so leverage on the remainder by spying on them.
Great business model eh?
The Reality Check
You cannot build a functioning system by strip-mining the physical world and ignoring the cost.
The tech bros built a god-like calculator, but they forgot they needed the people to actually feed it real meaning. They isolated themselves, automated away their own purpose, and alienated the entire publishing world.
The house is actively burning. Can Google change course before it's too late?
No one knows, do one? - Fats Waller

No comments:
Post a Comment