Michael Mauboussin analyzes firms’ growth forecasts tied to their investment in AI. He explains that base rates for U.S. public companies over the past 75 years suggest that OpenAI and Oracle Cloud have a low probability of achieving their five-year revenue projections.
Bayes and Base Rates
Counterpoint Global
Michael Mauboussin
Research
12 Pages
Key Takeaways
Base rates first: Start with what usually happens, then adjust beliefs as new evidence arrives.
Forecasts need context: Big growth targets should be judged against history, not enthusiasm or anecdotes.
Value not growth: Even rapid adoption can disappoint if competition and reinvestment needs erase economic returns.