We're all unpaid trainers for AI companies and we don't even realize it
Raw AI output still reads like AI. We all know the taste by now. But here's the thing: when someone who actually knows how to write takes that output and reworks it, or when a skilled writer uses AI as a tool to speed up their own process, the result is genuinely decent. Sometimes even good.
Now think about what happens next. All of that refined content goes back onto the internet. Blog posts, articles, social media, newsletters. And guess what? It becomes training data for the next generation of models.
Every time you read a piece of AI content and think "this part feels off but that part is solid," you're doing data annotation in your head. You're comparing, evaluating, judging quality. And the platforms collecting engagement signals on that content are feeding exactly that information back to the models. What gets clicks, what gets shared, what gets ignored.
So in a way, every person writing with AI and every person reading AI content is an unpaid trainer for these model companies. The writers refine the output. The readers label it with their attention. The models scrape whatever performed well and use it to evolve.
I genuinely think these model companies should be paying us at this point. We're literally training their models for free.