My AI team gets another pay rise.

My AI team has just had another pay rise. Up 20% in five months.

And I’m fairly sure the calculation still misses quite a few costs.

Even so, this is unquestionably one of the best investments I make every month. Even at ten times the cost, the ROI would still be fantastic.

Expensive is not the same as poor value.

But also: the era when the big AI companies lured us in with heavily subsidised prices is coming to an end. Soon it will matter more to understand which models should do what, when and at what price, rather than simply using AI.

So my advice is simple: experiment now. Compare. Learn. Build your AI team while experimenting is still cheap. A year from now, optimisation will probably be a competitive advantage.