Lesson Content

The problem with unstructured decisions.

Most people make decisions by listing pros and cons, then going with their gut. This approach systematically overweights recently encountered information, underweights unlikely-but-consequential risks, and fails to surface alternatives that were never considered.

Meta AI is a powerful tool for introducing structure – not to make decisions for you, but to ensure you are deciding with complete information.

The criteria-first method.

Before evaluating options, establish criteria:

"I need to make a decision about [topic]. Before I give you the options I am considering, help me identify: what are the most important criteria I should be evaluating this decision against? What am I likely to underweight because it is not immediately obvious? What criteria do people often regret ignoring for decisions like this?"

Only after establishing criteria:

"Now I am considering these options: [list]. Evaluate each against the criteria we identified. Be direct where one option clearly outperforms on an important criterion."

The reversibility test.

Before investing in a detailed decision analysis:

"Is this decision reversible or irreversible? What is the cost of reversing it if I make the wrong choice? Does the reversibility change how much analysis is appropriate?"

High reversibility -> decide quickly, learn by doing. Low reversibility -> invest in analysis before deciding.

The decision diagnostic.

For decisions made intuitively that feel right but you cannot fully explain:

"I have already decided to [decision]. I am asking you to be a critical reviewer: what are the strongest arguments against this decision? What assumptions am I making that might be wrong? What would need to be true for this decision to turn out badly?"

This is not to second-guess – it is to catch blind spots before it is too late to address them.

Weighted comparison.

For decisions with multiple criteria of unequal importance:

"I need to compare [option A] and [option B] on these criteria: [list]. The most important criterion for me is [criterion 1], followed by [criterion 2], and then [criterion 3]. Evaluate both options weighted by that priority and give me a clear recommendation."