Competitive models

Is it competitively advantageous for companies to use AI to do the same amount of stuff with less people, or do more stuff with the same people?

And if so, where does it lead? The model below is a starting point.

A vs. B

In this scenario, two identical firms implement the same AI in year zero, but adopt different strategies. Each firm earns a revenue of 1.00 a year and spend 0.60 on wages^?.

A: Same stuff, fewer people

Firm A uses the increased productivity of AI to produce the same stuff as before but with fewer people.

profit a year

B: More stuff, same people

Firm B uses the increased productivity of AI to produce more stuff with the same amount of people.

profit a year

Productivity multiplier

Value of new stuff

Death spiral or golden age?

In this scenario, we extend the above calculations to a societal level. How does the economy change if companies adopt strategy A or B? Does the economy take a death spiral or enter a golden age?

Proportion of companies choosing A vs B

Calculations & workings

Charting calculation

Each year, every company repeats its choice: A firms' cut wages reduce overall spending; B firms' new stuff adds to it^?.
New categories of stuff compound, unlocking the next^?.

Model defaults

Productivity multiplier = ×1.15. The best referenced study measured a productivity increase from AI of about 15%.^? This will likely change as technology matures.
Value of new stuff ×1.00. There is no evidence either way, so parity seems like a good starting point.
Strategy split 90:10. A study interviewed thirty executives, and found 91% of their generative-AI use cases automated worker tasks.^?