Value is created through innovation, but how much of that value accrues to the innovator depends partly on how quickly their competitors imitate the innovation. Innovators must deter competition to get some of the value they created. These ways of deterring competition are called, in various contexts, barriers to entry, sustainable competitive advantages, or, colloquially, moats. [Taxonomy of Moats - Reaction Wheel]
| Behavioral Bias | Description | Impact on Investment Decision-Making | Mauboussin's AI-Aided Countermeasures |
|---|---|---|---|
| Overconfidence | Overestimating one's knowledge or ability to predict outcomes. | Leads to excessive risk-taking, concentrated bets, ignoring contradictory evidence. | Consciously seek disconfirming evidence; consider a wider range of outcomes; temper optimism with base rates. Use AI tools to surface counterarguments and less obvious risks, generate scenario analysis beyond your initial views, and aggregate probabilistic forecasts from diverse sources to calibrate your confidence. |
| Anchoring / Tunnel Vision | Fixating on an initial piece of information or a narrow set of inputs. | Prevents objective reassessment of investments; leads to missing alternatives or holding underperforming assets. | Explicitly consider a full range of alternatives; seek dissent and opposing arguments; "red team" your own ideas. Leverage AI to generate alternative scenarios and viewpoints, identify overlooked opportunities or risks, and simulate decision processes from the perspectives of other investors or market participants. |
| Hindsight Bias | Believing, after an event, that one had predicted or known the outcome all along. | Impedes accurate learning from past decisions; fosters false confidence. | Keep a decision journal, documenting rationale, information, and expectations before decisions are made. Employ AI to timestamp, catalog, and later analyze the original logic and context of decisions, ensuring post-outcome reflections are anchored to what was actually known and believed at the time. |
| Emotional Extremes | Making decisions under heightened emotional states (greed, fear, euphoria, stress). | Rapid erosion of decision-making skills; leads to irrational choices. | Postpone important decisions when emotionally charged; seek emotional poise. Use AI sentiment tracking to identify periods of heightened emotional language or erratic behavior in your records; set up automated reminders or restrictions to prevent trading or major decisions during high-stress periods. |
| Misunderstanding Incentives | Failing to recognize how incentives (financial/non-financial) shape behavior. | Leads to misjudging motivations of market participants or management. | Carefully consider all existing incentives and the behaviors they might motivate, including subconscious effects. Utilize AI to map and analyze incentive structures for all stakeholders, uncovering potential misalignments or hidden motivations that might influence outcomes in ways not readily apparent. |
| Overreliance on Intuition | Depending on "gut feelings" in complex, unpredictable environments. | Can lead to disastrous decisions in non-linear systems; fosters false sense of expertise. | Recognize when intuition is appropriate vs. when analytical rigor (System 2 thinking) is needed. |
| Base-Rate Neglect | Ignoring statistical averages in favor of specific, vivid information. | Leads to overly optimistic or pessimistic predictions; misinterpreting probabilities. | Always refer to base rates (reference classes) for more realistic probability estimates. Use AI to access, aggregate, and analyze large datasets to establish accurate base rates and reference classes, ensuring decisions are grounded in objective historical probabilities rather than anecdote or vivid detail. |
"Steve Jobs would have never shipped this."
"No way would this happen if Steve Jobs were still alive and running Apple."
Oh yeah? Which Steve Jobs are you referring to?
We have to learn the bitter lesson that building in how we think we think does not work in the long run. The bitter lesson is based on the historical observations that
- AI researchers have often tried to build knowledge into their agents
- this always helps in the short term, and is personally satisfying to the researcher, but
- in the long run it plateaus and even inhibits further progress, and
- breakthrough progress eventually arrives by an opposing approach based on scaling computation by search and learning.
The eventual success is tinged with bitterness, and often incompletely digested, because it is success over a favored, human-centric approach.
I think about this a lot right now re: Waymo v. Tesla.
.jpg)
