When predicting major economic shifts like a bond market crisis or an AI stock correction, being wrong in a specific year doesn't invalidate the thesis. The underlying pressures may still exist, with the predicted event simply postponed. This reframes forecast misses as primarily errors in timing rather than analysis.
Related Insights
While AI technology will achieve widespread adoption and major breakthroughs, the financial infrastructure supporting it will falter. Peripheral companies that jumped on the AI trend without a core business will face a significant market correction, creating a paradoxical "best and worst" year for the industry.
The feeling of correctly predicting economic disaster due to flawed policies, yet being powerless to stop it, is akin to the Greek myth of Cassandra. She was cursed to know the future but have no one believe her. This illustrates the frustration of seeing knowable, mechanistic failures unfold while warnings go unheeded.
A 2022 study by the Forecasting Research Institute has been reviewed, revealing that top forecasters and AI experts significantly underestimated AI advancements. They assigned single-digit odds to breakthroughs that occurred within two years, proving we are consistently behind the curve in our predictions.
The market for financial forecasts is driven by a psychological need to reduce uncertainty, not a demand for accuracy. Pundits who offer confident, black-and-white predictions thrive because they soothe this anxiety. This is why the industry persists despite a terrible track record; it's selling a feeling, not a result.
![[REPLAY] Morgan Housel – Same as Ever (EP.359) thumbnail](https://static.libsyn.com/p/assets/f/b/7/6/fb76e1c8bfb69e8dd959afa2a1bf1c87/CA_Square_logo_light_background_6.26.24.jpg){kind=logo}
History is filled with leading scientists being wildly wrong about the timing of their own breakthroughs. Enrico Fermi thought nuclear piles were 50 years away just two years before he built one. This unreliability means any specific AGI timeline should be distrusted.
Long-term economic predictions are largely useless for trading because market dynamics are short-term. The real value lies in daily or weekly portfolio adjustments and risk management, which are uncorrelated with year-long forecasts.
A leading AI expert, Paul Roetzer, reflects that in 2016 he wrongly predicted rapid, widespread AI adoption by 2020. He was wrong about the timeline but found he had actually underestimated AI's eventual transformative effect on business, society, and the economy.
Howard Marks highlights a critical paradox for investors and forecasters: a correct prediction that materializes too late is functionally the same as an incorrect one. This implies that timing is as crucial as the thesis itself, requiring a willingness to look wrong in the short term.
Michael Mauboussin's BIN framework reveals that inconsistent judgments ('noise') are often a larger source of forecasting errors than personal biases or insufficient information. Reducing this variability through methods like combining independent judgments is a key to better decision-making.
Investors often believe their analysis is correct even if their timing is off, leading to losses. The reality is that in markets, timing is not a separate variable; it's integral to being right. A poorly timed but eventually correct bet still results in a total loss.
