Doesn't sound like we can support what is coming..
If it follows the normal technology life cycle, I expect it to follow something like Moore's Law.
Most of what is in a LLM is garbage. It works, but it is hugely expensive. As models refine, the garbage is continually sorted each generation. That is core to what DeepSeek did. They further refined the garbagy models of other companies. But those companies just stole content off the internet so cry me a river.
Eventually AI will be running locally in your phone using tiny fractions of the energy.
So it is risky to try and project in the future based on the current state of the tech. Tech isn't static.
This was a mobile phone in the early 80's:
I expect the capability and efficiency of AI to improve exponentially and demand to increase exponentially. Hopefully similar curves.
This is Capitalism. It will use up all the recourses the Market can bear. As long as companies believe it will make them more than it costs. I expect some new equilibrium to be achieved.
Now, whether Capitalism still makes sense after AI gets big enough, is a more complex question. I suspect there will be some reorganization of society. It will be unavoidable if you have 50% unemployment among educated knowledge workers. Those people will vote. Some accommodation is going to be made.
But you never offered a remedy. Do you think we should pass laws outlawing AI?
