ReflectionPerspective

Training the machine

AI once learned from books and science. Now it trains on our noise. Maybe progress isn’t about how much it learns, but where lessons come from.

When AI first arrived, it felt like magic. It could draw from centuries of books, journals, and science. The best of what we’d written down.

Now it’s trained on itself. On our posts, our prompts, our throwaway comments. Each new layer built from the noise of the crowd.

That raises a question: are we making it smarter or dumber?

Human progress came from the sharp edges of knowledge. From people who pushed fields forward, not from a blend of everyone’s chatter.

If AI is going to change the world for the better, maybe the real work isn’t scale.

Maybe it’s not about how much AI learns, but where the lessons come from.

Training the machine { Reflections | Liam Barr }