Books I read in 2022

In 2022, I set myself the goal of reading one book a quarter — a low goal but something I felt I could commit to. I ended up reading nine books over the year, of which I’ve provided summaries and brief thoughts on seven below.

The other two books I read were Between the World and Me by Ta-Nehisi Coates and My Time Among the Whites: Notes from an Unfinished Education by Jennine Capó Crucet, both of which deal with themes of racism and otherness in American society. I enjoyed both but didn’t feel that they were books that I’d necessarily refer back to later, so I didn’t make notes for them.

The notes below attempt to distill the key points of each book into useful snippets. Omissions and errors are my own.

Getting to Yes

This is a surprisingly slim book with deep insights that reflect decades of work in the field by the authors: one of the highest wisdom-to-page ratio books I’ve ever read. Some of the examples from the 1960s and 1970s now feel a bit dated, and reflect the authors’ experience in international diplomacy and labor negotiations (areas far from the typical concerns of a reader in the tech industry), but the advice is stellar and worth thinking about deeply. I highly recommend this book.

Book Summary

Most negotiations take place via positional bargaining: people take positions and make concessions to reach agreement. This type of negotiation is inefficient, can damage relationships, and may not even be possible when there are many parties involved in the negotiation.

Instead, the authors advocate for principled negotiation, based on the following principles:

• People: Separate the people from the problem
• Interests: Focus on interests, not positions
• Options: Invent multiple options for mutual gain before making decisions
• Criteria: Use objective criteria to evaluate which options to choose

Separate the people from the problem:

• Negotiators have an interest in both the specific negotiation in question as well as in the long term relationship. You can disentangle the substance of the negotiation from the relationship itself.
• Understand the other side’s perceptions, and clearly express your own perceptions. [This concept is similar to “adding to the pool of meaning” from Crucial Conversations.]
• Understand the other side’s emotions; legitimize them by discussing them explicitly.
• In conversations, listen actively and acknowledge what the other side is saying, paraphrasing what they say. Then state your own ideas clearly. [This is “seek first to understand, then to be understood” from The Seven Habits of Highly Effective People, and also echoes advice in Crucial Conversations.]
• Try to build long-term relationships with people.
• Face the problem, not the people — this can be done by literally sitting on the same side of a table, with documents of the problem in front of you.

Focus on interests, not positions:

• Try to understand what the other side truly wants, which they may be expressing as a position but which may actually be satisfied by another option.
• Articulate your own interests clearly, and acknowledge the other side’s interests.
• Look forward rather than assigning blame for what happened in the past.
• Be hard on the problem and soft on the people.

Invent options for mutual gain:

• Brainstorm solutions without making any decisions. You can do this effectively if you understand each other’s interests.
• Identify the most promising ideas and try to improve them.
• To come up with good options during brainstorming, use the “Circle Chart” method: Problem (what’s wrong?) -> Analysis (understand the causes) -> Approaches (broad ideas about what to do) -> Action Ideas (specific steps).
• Identify common and complementary interests that may allow you to invent options that satisfy all parties.
• Identify options that will be easy for the other side to agree to, and identify who needs to make the decision.

Insist on using objective criteria:

• Objective criteria are necessary to evaluate the options generated in the previous step, to avoid the negotiation becoming a battle of wills.
• Frame issues as a joint search for objective criteria. Be flexible about which standards you use, but insist on using principles rather than pressure to make decisions about options.

For negotiators who are in a more powerful position than you, develop your BATNA — your best alternative to a negotiated agreement.

If the other side won’t play along, you can use what the authors called negotiation jujitsu (never responding directly to attacks, but instead deflecting them with questions) or call in a third party to come up with a single proposal based on ideas from both parties (the one-text procedure).

With hard bargainers, you should negotiate about the “rules of the game”, or the form of the negotiation itself. Explicitly raise the fact that they may be using hardball tactics, and talk about how you can change that. You can use the same principles as above (separating the people from the problem, focusing on interests rather than positions, inventing options for mutual gain, and insisting on using objective criteria).

Range

I read this book because it was a Bill Gates recommended book in 2020. This is a popular science book: it’s entertaining and provides many anecdotes, but it’s too long and makes its case multiple times in slightly different ways. It could have been much shorter and had the same impact.

Book Summary

Repetitive practice leads to expertise only in certain kinds of situations, or learning environments. There are two types of learning environments: kind learning environments, in which patterns repeat over and over, and feedback is accurate and rapid (for example, golf or chess); and wicked learning environments, in which patterns don’t repeat, and feedback is often obscured or misleading (for example, predicting job performance or assessing student potential).

People with diverse experiences can weave those experiences together to come with unique insights. For example, Nobel laureates are 22x more likely to have deep artistic interests than the general population of scientists. Another example is a company called InnoCentive, which runs contests to solve challenges from a variety of disciplines (for example, cleaning up the Exxon Valdez oil spill) that are often solved by laypeople rather than specialists in a field.

In some cases, specialists in an area perform poorly even when making predictions within their area of expertise — for example, financial analysts. The author distinguishes between “hedgehogs”, who have narrow specialization in one area, and “foxes”, who have broad experience in many areas. Foxes tend to make better predictions in wicked learning environments. Groups of foxes can become very effective “super-forecasters”.

Education should be focused on conceptual, transferable knowledge rather than specialization — the opposite of what most schools and universities do.

“Spacing” or “interleaving” — practicing the same material days, weeks or months apart — is a good technique for retaining concepts long-term. Analogies are a good tool for reasoning about things that you don’t have direct experience with (for example, Kepler used analogies to terrestrial processes to come up with his laws of planetary motion), and broad experience in a range of areas can help in coming with those analogies.

Matching yourself to the right opportunity (“match quality”) is important to do good work, and a broad range of experiences can help you sample many different areas before deciding which one is a good fit. Persisting in something that isn’t a good fit is not a good idea.

In life, the best approach may be to simply plan for the short term, and take good opportunities as they come up (“test-and-iterate”). Paul Graham calls this “working forward from promising situations” as opposed to “working backward from a goal”.

An enthusiastic, childish playful streak is a recurring theme in research on creative thinkers.

Onward

This is the story of how Howard Schultz turned around Starbucks during the 2008 “Great Recession”. While I did not take away many transferable lessons from the book, given that the events it narrates are highly specific, I found it a fascinating look into what happened at the highest levels of a major company during a critical time. Howard Schultz comes across as a lot more ruthless than his benevolent public persona (as he may have needed to be during a difficult time).

Top takeaways

Growth without discipline can be very dangerous to a company. Starbucks had expanded very rapidly in the 2000s, but many of those stores were not profitable, and led to a dilution of the Starbucks experience.

Howard Schultz had a clear idea of where he wanted to take the company, though he did “launch and iterate” along the way, with some failures. He was not reluctant to let people go (for example, the Starbucks CEO) to bring the company out of its decline.

Process for turning around the company:

• Come up with seven pillars under which to group various initiatives (the “Transformation Agenda”) in collaboration with a small working group
• Change top leadership where necessary
• Invest in operational excellence such as supply chains, IT systems etc
• Conduct a summit for 200 top leaders and a separate conference with 10,000 managers to align them with the Transformation Agenda
• Invest in innovation initiatives (e.g. “Pike Place Blend”, Clover coffee machines)
• Expand into new areas (VIA instant coffee)
• Talk about the transformation agenda at shareholders meetings

Crucial Conversations

This is a spectacular book worth re-reading many times: it lives up to the hype. While people may organically develop some of the techniques described in this book (I could recognize some of the things I do that map to advice in the book), this was the first time that I had seen such a systematic framework for important conversations laid out.

The book is detailed, with topics, sub-topics and sub-sub-topics. Some of this complexity can feel excessive, but none of it feels redundant — these are genuinely deep, nuanced concepts. It took me a while to read this book just because of the depth of information in it: I re-read portions, did the exercises, and tried to internalize the information. I highly recommend this book.

Book Summary

A “crucial conversation” is one that has all of the following attributes: there are opposing opinions, emotions run strong, and the stakes are high. Most people handle crucial conversations poorly because of the ingrained “fight or flight” response.

Most people believe that when faced with a crucial conversation, they have one of two choices: (1) speak up and ruin their relationship with the other person, or (2) stay silent and let a bad situation continue. The authors call this the “Fool’s Choice”, and people who are good at crucial conversation take a third way: to be honest and respectful.

To have a crucial conversation, you must bring people to dialog, which involves filling the “pool of shared meaning” with everyone’s opinions and ideas, even if they appear wrong or controversial. People good at crucial conversations are also good at enabling all participants to fill this pool of shared meaning.

The following are the steps in crucial conversations.

Step 1: Start with Heart — set the foundation for a good crucial conversation

Figure out what you really want: for yourself, for others, and for your relationship with other people. Ask yourself whether you are behaving consistently with those desires.

Refuse the Fool’s Choice, and look for ways to bring everyone to dialog without being disrespectful or staying silent.

Step 2: Learn to Look — identify how things are going at the beginnings of a crucial conversation

Identify when a conversation turns crucial. You can detect this by looking at signs within yourself: physical signs such as your stomach getting tight, emotional signs such as feeling scared or angry, or behavioral signs like raising your voice.

When you detect that you’re in a crucial conversation, make sure to maintain safety, where people feel safe saying what they really feel. Do not respond to aggressive or inappropriate behavior from others, which may be a sign that they are feeling unsafe: instead, aim to bring the conversation back to safety.

When people feel unsafe, they resort to either silence or violence:

• Silence means withholding information from the pool of shared meaning. The three types of silence are: (1) Masking: Obscuring what you really mean through sugarcoating or sarcasm; (2) Avoiding: Steering the conversation away from sensitive topics; (3) Withdrawing: Pulling out of the conversation entirely
• Violence is when you try to compel or force your point of view on someone. The three most common forms of violence are: (1) Controlling: Forcing others to your point of view through overstating your facts or cutting other people off; (2) Labeling: Putting a negative label on other people’s ideas to discredit them; (3) Attacking: Directly threatening or belittling another person

When you’re in a crucial conversation, you need to monitor your own behavior for signs of silence or violence, as well as that of others.

Step 3: Make it safe — what to do to restore safety when you realize that people are feeling unsafe

In order to establish safety in the conversation, you must have both of the following:

• Mutual Purpose: Others should perceive that you are working toward a common goal, and that you care about them. If you feel like you and others don’t have mutual purpose, you can use the "Start with Heart" questions (what do I want for myself, for others, for the relationship) to bring mutuality into the conversation
• Mutual Respect: Others should feel that you respect them, or the conversation will become based on emotions such as fear and anger.

If you see that either mutual purpose or mutual respect are at risk, you should “step out” of the conversation by doing one or more of the following:

• Apologize: Offer a sincere apology for something you may have done wrong
• Contrast: Clarify that you don’t intend to disrespect the other person, and clarify your real purpose as a contrast to what the other person might be thinking is your intent. (“I didn’t mean to come across as not valuing your work. I think your work has been amazing. I am trying to help you get even better.”)

If you do not naturally have a mutual purpose with the others in the conversation, you can work to create a mutual purpose using the “CRIB” method, described below. For example, in a situation where two teams have different preferences about working weekends to achieve a company goal, you might use these skills as follows:

• Committing to seek a mutual purpose: “Arguing isn’t helping. Why don’t we see if we can come up with something that satisfies everyone?”
• Recognize the deeper purpose: Understand why people might not want to come in on a specific day or stay late
• Invent a mutual purpose: “Let’s make sure that our goal is that our working relationship stays positive and collaborative, regardless of how we feel about this specific issue.”
• Brainstorm strategies toward that purpose: Figure out ways for teams to work around each other’s schedules, but now hopefully in a more collaborative, less angry atmosphere.

Step 4: Master my stories — how to get your emotions under control during a crucial conversation

Between what happens to you or what you observe, and your actions in response, are the stories you tell yourself. These stories in turn create feelings that then result in actions. The authors call this your Path to Action: See and hear -> Tell a story -> Feel -> Act.

Learn to understand and change your stories by working backward from your actions (“retracing your Path to Action”). First, notice your behavior and identify whether your behavior has signs of “silence” or “violence”, and ask what you are feeling that’s making you act that way, and in turn what stories you are telling yourself that could be making you feel that way. Finally, try to understand what evidence may be supporting those stories, separating facts from interpretation.

Specific examples of stories you may be telling yourself:

• Victim Stories: You may feel that you are an innocent victim.
• Villain Stories: You may exaggerate someone else’s bad points and make them out as the sole person to blame for a situation.
• Helpless Stories: You may convince yourself that there’s nothing you can do but the sub-optimal actions you’re taking.

Once you understand the stories you’re telling yourself, you can convert them into more constructive stories by trying to consider whether you had a role in getting the situation to where it is, and trying to understand why a rational, decent person would do the things that you’re seeing them do.

Step 5: STATE my path — explain your thinking clearly and compassionately

You can use the “STATE” skills to talk about delicate topics:

• Share your facts: Start with the facts that are leading you to have the crucial conversation, not with your interpretation (your stories, which you have separated from the facts in the previous step). For example: “I notice that you have asked me for more detailed progress reports than other people on our team”.
• Tell your story: Talk gently but confidently about the conclusions you’ve come to from the facts (“I’m wondering if you’re having trouble trusting me to do my job”). You can use contrasting to soften your message and rebuild safety.
• Ask for others’ paths: Ask other people to share their own reasons for the behavior you’ve observed.
• Talk tentatively: Couch your observations in tentative language (“I’m wondering if…”) so that you give others the space to share their points of view.
• Encourage testing: Genuinely encourage others to share their point of view — this is especially important when you are in a position of power over the people you’re having a conversation with. You may want to play the devil’s advocate, and argue against yourself, to make people see that it’s ok to offer a different point of view.

Step 6: Explore others’ paths — deeply understand others’ points of view

You must truly want to understand what the other person is feeling. You can then use the “AMPP” listening skills to explore other people’s paths:

• Ask: Ask them to share their paths (points of view)
• Mirror: If you feel that the other person is not opening up, tell them how it appears to you that there is more under the surface
• Paraphrase: Repeat what they’ve shared with you to validate your understanding (and confirm that you’ve been listening carefully)
• Prime: In cases where it’s hard to get the other person to open, you can start the conversation by stating a hypothesis for what their point of view might be (“Are you feeling…”)

Once you’ve heard the other person’s point of view, you can use the “ABC” skills to provide a synthesis of your and the other person’s points of view:

• Agree: Say that you agree with the areas that you do agree with
• Build: Fill in gaps where something may have been left out, but do so in a constructive rather than destructive way (rather than saying “You’re wrong, you forgot about…”, you can say “Totally agree — in addition, I noticed that…”)
• Compare: Where there is a genuine disagreement, say so explicitly, but without implying that the other person is wrong. You can use the “STATE” skills from the earlier section to clearly but respectfully articulate your point of view.

Understanding other people’s points of view deeply can help you reach common ground with them.

Step 7: Move to action — move beyond the crucial conversation to make decisions and turn them into action

Through the preceding steps, you can help fill the “pool of meaning”. Once it’s filled and you’ve had the crucial conversation, you should move to decisions and action.

To make decisions, use one of the following styles: command (a single person makes the decision); consult (gather information from a group and let a subset of the group decide); vote; or consensus (very time-intensive, and should only be used for very complex issues where you absolutely need buy-in from everyone).

After you’ve made the decision, make sure to also decide “who will do what by when”: assign a specific person to create specific deliverables by a certain date. Agree on how you will check in on progress toward those deliverables.

The two most effective levers for having a crucial conversation are "Learning to Look" (identifying when things are going off the rails) and "Making it Safe" (showing concern for the other person to help them participate fully in the conversation).

Bad Blood

This is a very entertaining account of the fraud at Theranos by the journalist who exposed it, written in the style of a fiction thriller (except that it’s not fiction). The writing is sharp, though it sometimes dives too deep into tangents, and it’s possible to read the book over a weekend.

Top takeaways

Theranos was founded in 2003 and shut down in 2018, a 15 year period. The first Wall Street Journal article exposing the fraud came out in 2015. It was able to run a scam for an incredibly long time.

The company focused on a lot of activities ancillary to the product — designing the case for the device using ex-Apple designers; tie-ups with Walgreens, Safeway and other partners; an advertising campaign through TBWA; press coverage in Fortune, The New Yorker etc — all while the very core, the testing technology itself, did not work. It was an example of pre-product-market-fit scaling, something that it had in common with many startups. The difference was that most startups don’t engage in outright fraud in the way that Theranos did.

The company churned through a lot of people, but seems to have been able to continue hiring talent with good backgrounds till fairly late. Until news of the fraud broke, it didn’t seem like their poor treatment of employees had much of an effect on them being able to recruit good talent.

Timeline of exposé:

• Adam Clapper (a pathologist) published an article on his blog about being skeptical of Theranos after reading a New Yorker article about Elizabeth Holmes
• John Fuisz (former Holmes neighbor and long-time enemy of Holmes) read the blog post and put Clapper in touch with Phyllis Gardner (Stanford biology professor) and Rochelle Gibbons (widow of Ian Gibbons, scientist at Theranos who had committed suicide)
• Alan Beam (pathologist at Theranos) looked up John Fuisz’s LinkedIn profile (the book does not say why), Fuisz noticed it, contacted him, and got him in touch with Clapper
• Clapper contacted Carreyrou who started the investigation

Takeaway: what if Elizabeth hadn’t made a determined enemy in Fuisz? What if Fuisz hadn’t come across Clapper, or Clapper didn’t know Carreyrou, or Carreyrou wasn’t as good a reporter as he was? Sometimes, things happen in certain ways due to a series of fortuitous occurrences. It’s possible the company would have still been exposed in some other way, even if this specific chain of events was unlikely.

Theranos put a lot of legal pressure on the Wall Street Journal, and some mildly underhanded tactics to intimidate witnesses, but nothing ever devolved to actual violence.

Without an actual government agency (CMS, the federal department that oversees labs) taking a complaint from Erica Cheung (a Theranos chemist) seriously, it could be that the newspaper article would not have had any effect. It needed someone with enforcement authority to act.

What Got You Here Won’t Get You There

I found this book shallower than its reputation had made me believe. The author lists twenty habits that hold people back, but I found myself struggling to differentiate between some of them. The steps the author provides for identifying the specific habits that are holding you back, and for then changing them, seem overly simplistic.

This is a good quick read, but many of the principles could probably be condensed into a few denser ones.

Book Summary

Successful people can have habits that prevent them from becoming even more successful. These are largely habits in the area of interpersonal relationships; the higher up you go, the more your problems are behavioral rather than skills based.

The twenty most common habits that hold people back are:

• The need to win at all costs
• The desire to offer an opinion on every idea that someone else has
• The need to judge others
• Making destructive or sarcastic remarks
• Overusing negative qualifiers like “no”, “but”, or “however” when responding to someone (thereby telling them you disagree with them)
• The need to prove how smart they are: for example, telling people “I already knew that” in response to an idea or suggestion
• Being angry in a professional environment; it’s especially hard to overcome a reputation for emotional volatility
• Too much negativity, or telling other people why their ideas won’t work
• Refusing to share information
• Being unable to praise or recognize others
• Claiming credit that they don’t deserve
• Making excuses for their behavior, rather than just saying “sorry”
• Blaming their behavior on past events or people
• Failing to treat people fairly, or playing favorites
• Refusing to apologize
• Not listening to others
• Failing to thank people
• Punishing the messenger, or retaliating against people who give them bad news
• Passing blame onto someone else
• Believing that their faults are “really who they are” rather than flaws to be fixed

Getting feedback helps you understand which flaws are the most important to fix. Most people have flaws, but only some flaws are large enough that you should work on them.

Take the following steps to improve:

• Apologize: tell everyone who will listen that you are sorry for your past behavior
• Advertise: tell people that you will improve, and specifically what you will improve on
• Follow up with people periodically by asking them if you are improving. Listen to them and thank them for their comments.
• Practice “feed forward”: a technique where you ask people for suggestions on how you can improve in the future in the area you’ve picked

Thinking, Fast and Slow

This is a richly detailed book about the workings of the human mind. Like with Getting to Yes, the content here reflects a lifetime’s worth of work: the author is one of the founders of the field of behavioral economics, many concepts from which are major topics in this book.

The book is large (500 pages) and dense with concepts, so the notes below are very high-level. It’s a worthwhile endeavor to read the entire book, though it requires a significant time investment, and some of the insights may feel like amusing bits of knowledge rather than a systematic framework. Regardless, the book made me much more conscious of the sort of biases that all humans are prone to, which is valuable.

Book Summary

Key Concepts:

• Two systems: System 1 (intuitive, quick, automatic) and System 2 (effortful, systematic)
• Two models of people: Econs (rational economic agents) and Humans (people are they really are)
• Two selves: the experiencing self (experiences events as they occur) and the remembering self (remembers events later)

System 1 and System 2:

Our mind consists of two systems:

• System 1 operates automatically, quickly, and without voluntary control: for example, answering “what is 2+2” or detecting where a sound is coming from.
• System 2 performs complex mental computations that require effort: for example, multiplying large numbers together, or looking for a specific person in a crowd.

There are several tricks that can be used to make System 1 believe something; it’s useful to use System 2 to examine whether the judgements of System 1 are correct, at the cost of some speed and effort.

System 1 uses the “halo effect” to construct a coherent mental model of a person, filling in traits that you don’t have any information for e.g. if you like someone, you will consider them more generous even without any direct evidence of their generosity.

This can lead to biases where final grades for students on a paper are disproportionately affected by their answers to the first few questions; the halo effect ensures that later questions are more leniently graded.

The halo effect leads to jumping to conclusions on the basis of limited evidence, which the author calls “WYSIATI”: what you see is all there is, the tendency of System 1 to assume that it can construct a coherent model of a person or situation without stopping to consider whether there is other missing information that might be contradictory.

Humans and Econs:

Humans use heuristics and have biases, as opposed to the perfectly rational agents assumed by classical economics. A large part of the book explores these human heuristics and biases.

Humans use three types of heuristics when asked to answer questions that require estimations or judgement, each of which can result in incorrect answers.

The representativeness heuristic is used for questions like “What is the probability that object A belongs to class B”. This heuristic estimates a probability based on how representative A is of B, neglecting broader information, such as “base-rate frequency” (the prior probability of object A belonging to class B given no information about A).

For example, given a description such as “Steve is shy and withdrawn, and has a passion for order and detail”, and asked to predict whether Steve is more likely to be a librarian or a farmer, people will guess that Steve is a librarian, even though the proportion of farmers in the general population is much higher than librarians.

The “base-rate frequency” in this example would be the probability of a random person being a farmer vs a librarian, given no information about the person.

The availability heuristic is used when estimating the probability of an event or the frequency of an occurrence; it assigns a higher probability to events that more easily come to mind.

For example, people were given a list of well-known personalities of both sexes, and asked to estimate whether the male or female names were more common. If the male names were of more famous people than the female names, people were more likely to estimate that male names were more common, because they were easier to recall; and vice versa if the female names were of more famous people than the male names.

The adjustment and anchoring heuristic is used to make estimates, by starting from an initial “anchor” value, and then adjusting it.

The anchor value may be suggested as part of the problem, or by a partial computation. For example, groups asked to estimate the product 8x7x6x5x4x3x2x1 estimated a much higher value than groups asked to estimate the product 1x2x3x4x5x6x7x8, because the product of the first few numbers in the first formulation was much higher than in the second.

Prospect Theory:

The author developed a model called prospect theory that describes people’s behavior around choices that involve gains and losses (what the author calls gambles). The key assertions of prospect theory are:

• When making a gamble, relative values are more important than absolute values — for example, the difference in subjective value between a loss of $200 and a loss of $100 is greater than that between a loss of $1200 and a loss of $1100.
• People feel more disappointed by a loss of $X than they feel happy with a gain of the same amount.
• People are more inclined to try to take risks to avert losses rather than take risks to get gains — called loss aversion. In other words, people are risk averse around gains (they prefer a lower gain with lower risk) and risk seeking around losses (they prefer to take a risk with a small probability of reducing the value of a sure loss).

Framing a problem in a way that triggers loss aversion can produce seemingly irrational choices e.g. when asked to choose between two modes of treating an epidemic, one of which is presented as saving N lives and the other resulting in M deaths, people will choose the former option even when the two options are mathematically identical.

People tend to make judgements about others based on very limited information that may have a limited correlation with the attribute they are evaluating. For example, if people are asked to predict a college senior’s GPA given that they were a good reader at age 4, they will typically predict a much higher value than the correlation between early reading ability and college GPA would suggest. The book offers a way to systematically eliminate this bias:

• First, use the base-rate frequency to come up with a value for the desired attribute. In the example above, this might be the average GPA of all college seniors.
• Next, use the information about the person provided to come up with an intuitive estimate for the desired value (the original GPA that you might have thought of)
• Estimate how complete the information you have is, and how much it correlates with the desired value (how much correlation is there between reading early and college GPA?)
• Adjust your base estimate in the direction of your intuitive prediction by the amount of correlation — that is, if you think that the correlation is 0.3, move from your initial estimate based on base-rate frequency to your intuitive prediction by 30%.

How should we determine when to trust an expert’s intuitive judgements? People are often over-confident about their judgements, even experts, and use information that’s weakly correlated with what they are trying to predict. We should look for the following factors:

• An environment that’s regular enough to be predictable
• A feedback loop that lets the person making the predictions learn those regularities through practice [This is the concept of “kind” vs “wicked” learning environments from the book Range, discussed earlier.]

Experiencing and Remembering Selves:

All people have two selves: the experiencing self, that experiences events while they are happening, and the remembering self, that holds the memory of events after they are complete.

The remembering self usually dominates; people use their memories of experiences to make decisions about what to do in the future.

The remembering self makes memories based on the following:

• Peak-end effect: The average of the peak experience and the experience at the end of the event
• Duration neglect: Memories are unaffected by the duration of the event.

That is, people prefer an operation that lasts longer but has a lower peak amount of pain and ends with less pain than a shorter operation that may have a higher peak level of pain. Their memory of the experience is not the “area under the curve” of the total pain experienced during the operation.

This article also appears on my blog.