Billions of workers make their living selling intelligent labor. 

AI can now perform SOME PORTION of that intelligent labor more cheaply and more efficiently than most humans. That portion is growing quickly.

If you can identify a tiny niche where AI is more efficient than humans, you can arbitrage the difference between what humans earn (say $100 per unit) and what AIs "earn" for the same product (say $1 per unit).

100x return on AI spent (ROAIS) is possible - even likely in many industries.

"So is paying $200/m for an AI too expensive? What is the upper limit of spending for AI?"

The answer is simple: there is no ceiling. You want to spend as much money as you can on true AI arbitrage opportunities.

And you can scale up fast: (example)

• Step 1: Spend $10 for AI tokens to create 10 intelligent labor units earning you $1000.
• Step 2: Reinvest these $1000 into AI tokens to create 1000 intelligent labor units earning you $100,000.

And keep scaling up. 

Given this first principles reasoning, the only question we should be bothered with is: "Where can we use AI to automate a portion of human labor?"

Like in the Hitchhiker's Guide, it's all about the question. The answer is manifold - and much easier to figure out. Just look around.

ROAIS is the most profitable business opportunity in the world - and there is no ceiling: Intelligent labor is already the largest total addressable market (TAM) and it will keep expanding forever, with volatility, given infinite energy and entropy in the universe.

Be on the right side of change!