Center for Applied Rationality Workshop (2018)

“Do you know what you’re doing, and why you’re doing it?” According to Duncan Sabien of the Center for Applied Rationality, this is a key question to ask yourself throughout life. In this workshop from Effective Altruism Global 2018: San Francisco, he describes a few different techniques, including managing your personal autopilot and mimicking useful skills, that all rely on this core reflection. A transcript of Duncan's workshop is below, including input from the audience, which we have lightly edited for clarity.

My name is Duncan Sabien. I'm the curriculum director and COO at The Center for Applied Rationality in Berkeley. We put on four and a half day rationality workshops, and also some targeted programs for groups like AI researchers, mathematicians, so on and so forth. Our goal is to improve human rationality. In this talk, I'm going to try to give you a little taste of what we do at workshops, and what we've been figuring out, and what we're all about.

I'm sad to say that it's not going to work. There's an effect that we've noticed where when we give an hour long talk, people leave super excited about rationality, and about the potential to be better at achieving their goals and figuring things out. And what happens is, it's sort of like if you had just found out about the concept of pushups. You're like, "Man, pushups, that sounds so cool. I can be so strong, and my arms would look so good."

But there's a very big difference between learning about pushups and actually doing pushups. So to frame this talk, most of this talk will be me telling you guys about pushups, and how cool pushups are. But on base rates, very few of us will leave this talk and then go do pushups. So make it be you. You're going to be the one that breaks the frame, and goes out and actually does stuff. Make sense?

Cool. Occasionally there might be spots for people to ask questions and make comments. If so, we'll have a mic that we'll send around, because they're recording, and it makes it a lot easier for the recording to be sensible. Please talk into the mic. But otherwise, we're just going to dive right in.

My first question is, do you know what you are doing and why you are doing it? I'm actually going to repeat this question a whole bunch over the course of the hour. So get prepared to get bored by it. But right now you're in this room. I'd like you to take 20 seconds and check, do you actually know why you're in this room? What brought you here? Is there a reason? Are you just on autopilot? Take 20 seconds to check.

Every time I say take 20 seconds to check, you're also certainly welcome to just sit there not doing anything. I won't know. But from nope to yep, raise your hands to the extent that you know what you're doing and why you're doing it. So here is, "I have no clue." Here is, "I'm like a perfectly rational automaton." Nope to yep, can I get all the hands?

Neat. Not bad for EAG. Okay, so human rationality. My framing on human rationality is that it has two major parts. We're trying to get better at forming true beliefs, and letting go of false ones. And we're trying to get better at taking effective action. In other words, you want to actually be able to say the sentence, "It's going to work," and have that sentence be true. Right? And when you can't honestly say that sentence you should caveat, the way that I just caveated a second ago. So hopefully this is what you're interested in, because this is what we're going to dive into.

So, CFAR is an organization made up of specific people who have done specific research. CFAR has a canon. It's sort of an attempt to make rationality be a thing, and our canon is this large battery of techniques. I've drawn up a whole bunch of them. You can come take pictures of this at the end if you're interested. We have techniques with neat names like Double Crux, and Focusing, and Goal Factoring, and Systematization, and Resolve Cycles, and Murphy-Jitsu. And each one of these techniques is an attempt to solve a certain problem. I will say the problems a little bit slower so you guys can catch them.

We have a technique that we've invented to try to help navigate intellectual disagreement. We asked ourselves the question, why is it that some conversations, some fights go a hundred times better than other conversations and other fights. And we tried to techniqueify the difference. We have a technique that we stole, called Focusing, which is for drawing info out of your subconscious or implicit models. The idea is, how can we take advantage of all of the stuff that our subconscious model, like our system one that's listening all time, how can we take advantage of this information, that doesn't land in our verbal loop? How can we draw it out and get a handle on it?

How is it that sometimes you end up doing something that doesn't achieve your goals very well at all? The classic example is, raise your hand if you or someone you knew was going to law school or medical school because they wanted to make their parents happy? Is that the most cost efficient way to make your parents happy? So we have a technique that's designed to deal with mismatches between goals and plans. We have a technique for modeling just what's going on in your second-to-second cognition, what's actually going on in your brain.

We have a technique for bulletproofing plans. Somehow we have this capacity to make plans that just don't work. And have you ever confidently declared that you'll be at the airport by 7:00? Right? We have a technique for overcoming decision paralysis. We have a technique for automating away drains on your time and attention, for resolving internal conflicts.

The thing about all these techniques is they're all sort of fake, right? The art of human rationality is not evenly dividable into single kicks and punches. We make up these little kicks and punches so that people can practice the basics and ingrain them. But most of our staff members, most of our alumni rarely sit down to use goal factoring. We will, once in a while; the techniques are pretty cool. They do actually work.

But what's much more important is this holistic lump. This big, vague, misshapen, ill-defined lump of human rationality. The idea of being able to see what's real, and do what you're trying to do. So don't think of this as a checklist. Where like, "Once I've collected all the techniques, then I will be a superhuman rationality machine." In point of fact, people with that mentality tend to plateau very early on. What you want to do is keep looking.

All right. I will actually teach you some of the techniques, though. We're going to talk about TAPs today. And if you attended Eli's Double-Crux session yesterday, that was great. If you didn't, get a time machine. And later on today there's a Murphy-Jitsu session that will be pretty fun as well. But okay, this is not a workshop, and we only have an hour. So rather than trying to feed you all of that rationality in bite-sized chunks that won't work, what I want you to learn how to do is invent your own rationality... and that's what we'll spend the next little bit talking about.

By the way, do you know what you're doing and why you're doing it? Reactions of any kind are appreciated. Yes, okay. I'll explain later on why I'm going to keep repeating that point over and over again. But actually check, do you know what you're doing and why you're doing it? Okay, I'm going to give you three... those of you who can squint, you can see "Ah, there's three."

I'm going to give you three different lenses on how to invent your own rationality. This is basically how CFAR gets its material, where we learn things from. Sometimes there's something cool that comes up in the cog-science literature that we can just take. They're like, "Here's this new thing that we've discovered." And like, "Great, we can just teach that at the workshops." But very often cognitive science research doesn't really address the things that we care about, and so we have to cobble together our own rationality.

This is my favorite version. This is a timeline. You're on it right now. You're over here. Let's see if I have a blue in my pocket. Yes, this is you. And what's happening is you're on this timeline that's pointed toward the universe where you end up happy, and successful, and the good thing happens. But for some reason, over and over you end up down here.

This is like you discover that you have eaten an entire package of Oreos. Or you discover that you are once again having the same fight with your boss, or with your significant other. You discover that you are once again getting angry in traffic and your blood pressure is up. All of this sort of thing. This model make sense? There's a timeline.

All right, there are three places to look if you are trying to invent rationality, I claim. And those three places are, here, here, and here. Take a moment to stare at that. And if you are actively participating, form some opinions about why I circled those three spots. Nope to yep, do you have opinions? Neat, I hope they agree with mine, because that would be data that I may be right.

Okay, the idea here is humans are a process, a very complicated, messy lump of hundreds and hundreds of algorithms, all fighting at the same time. And somehow this process is ending up down here. I like to look in these three places. Because here, nothing has gone wrong yet. Here, we're just living our life. If you're the sort of person who has road rage, you haven't even gotten on the highway yet. If you're the sort of person who eats compulsively, you haven't even gone to the cabinet yet, right?

Your significant other walked in the door, and all you've said is, "Hi, how are you?" You haven't started talking about politics yet, right? So this is an interesting place to look, because there's something about the way that we move automatically through the world, kind of on autopilot. We'll talk more about our autopilot later in the hour.

But you're just like... and your autopilot is what takes you down here. If you've ever made a New Year's resolution and then by March you haven't made any progress on that New Year's resolution, because you literally just never thought of it again. That's your autopilot in action, right? The New Year's resolution would require you to do something different from what you normally do, and you just didn't. With me so far?

Cool. This point is also a fascinating point to look at. This is more obvious. This is where most people look. When did things go wrong? Right? Why did it all go wrong? Identifying this point is not always easy. If you're the sort of person who eats an entire package of Oreos, was this point the moment that you ate the fifth Oreo? Or was it the moment that you took Oreos down out of the cabinet? Or was it the moment that you purchased the Oreos?

It's not always easy to tell what the actual point divergence is. But if you can zero in on this point, and you can find the actual split where you went left and you should have gone right, according to your own values, you can build a lot of rationality out of that. Especially out of looking at it like, why did you go the wrong way in the first place?

There's a principle that we teach at CFAR. We loosely title it The Good Faith Principle. The idea is that everything that goes on in your brain, everything that you do should be taken in good faith. The idea is that every impulse that you have, every thought that floats across your consciousness is trying to be good, and trying to make the world good, and trying to make your life good.

The problem is sometimes it's just like, "Oops, mistaken." Like we really crave sugar, at least in part, because in the ancestral environment sugar was mostly found in things like fruit, which had a ton of calories, and had a ton of micro nutrients available. And so our bodies learned to like sugar, because that was a marker for other good things. And then we learned how to separate the sugar from the fruit, and like, "Oops," right?

But the part of you that craves sugar and is driving you towards sugar, and wants you to eat every single Oreo, it's trying to make your life good. It's trying to get you nutrients, and calories, and that sort of thing. It just happens to be miscalibrated for the world that you find yourself in. And so, rather than kicking the part of you that likes Oreos, it's better to like, "Good job, sugar craving. I see how you're on my side. You're trying, but we're going to go over this way where the broccoli is." So looking here and figuring out not just what went wrong, but why it went wrong is a fantastic place to unpack rationality.

Many of our techniques come from looking at that moment when things either go well or badly. And we look at people who do things... people who are dramatically effective. And we're like, "What was going on in their brain, second-by-second?" And we look at people who are dramatically ineffective, and like, "What was going on in their brain, second-by-second?" And we try to cross-compare and figure it out. Any questions so far? Yes, do we have the microphone? Yes.

Audience: Yes, so you didn't put a circle on the happy path, on the right there. Why not?

Yes, if you're already on the happy path, it's less urgent that you ask questions and figure out what's going on. It's still great to know what's going on even when things are going well. But it's more critical, I claim, to figure out what's going on when things are going badly.

The third circle, the reason for this third circle is, I like to look here more than most of my colleagues do. A lot of my colleagues are like, "We should focus our interventions here." But this is a place where I often have the benefit of noticing. I'll be sort of asleep, and then like, "Ye wake up, and ye find ye self having eaten two rows of Oreos, and there's one row left."

And it's like, "Well, I mean I've already fallen off my diet. I'm a terrible person, so I might as well." That's the failure with abandon kind of thing. I like to look in this place because my brain is often awake and aware at that point. Like if you are a road-rage type person, this is where... down here is where you and the other person have pulled off to the side of the road and are getting your respective baseball bats out of the trunk, right?

But here, this is the point where you've swerved in front of them and turned on your windshield wiper fluids so it's going to mess up their... it's bad. It's bad, but it's not hopeless. And it's a lot easier to correct back from here, than it is to get all the way back up from here. Make sense? So for me this is another primary place to look for interventions, and inventions, and figuring out, "Okay, how can I re-correct course from that spot?"

So, what we're going to do now, we're going to put 90 seconds on the clock. If you're exhausted, or you haven't woken up yet, feel free to just zone out. But if you're playing along at home, then you're going to take 90 seconds, and you're going to look for some dynamic in your life that matches this model that I've drawn here. Some place where you keep ending up in the frowny-face land.

And you're going to just take a quick, just quickly, quickly, see if you can find anything in any of these three spots. Make sense? Cool, 90 seconds, go. And If you can hear me, start humming... everybody humming. Social peer pressure to be quiet again... nice. That's how we do it CFAR workshops, so I don't have to shout all the time.

Cool, quick aside. You guys know the trope, where there's shoulder advisors. There's the angel on one side, and the demon on the other. This one was like, "Eat all the Oreos." And this one was like, "Eat your broccoli," yes? What I would like you to do, humor me for a moment, is pop back into this dynamic. Visualize it, if you have visual imagery. If you're the sort of person who doesn't have visual imagery, just do whatever it is that you do when you're reading a book.

But go back into that moment for me, and I want you to imagine a tiny little Duncan popping up on your shoulder. And what he says is, "Do you know what you're doing, and why you are doing it?" Do me a favor, actually visualize that. Cool, thanks. All right, our second method of inventing rationality. This is a method we call bow tie or hourglassing. We haven't actually turned it into a class yet, so we're not sure what the name's going to be. Naming things is very important.

But here we have a blue person, that's you, who's looking at an orange person who is awesome. The orange person is awesome at some skill. Maybe they're a really good writer. Maybe they're a really good athlete. And you, the blue person, want some of it. You want to invent some rationality and steal what they've got.

This method, by the way, comes in part from Ben Franklin, you can look up more about how Ben Franklin learned how to write. Basically, the idea was he would take articles that he admired from the papers, and he would stare at them really hard. And he would develop theories about the principles that the author of the article was adhering to. So you take a large corpus of data, and you try to boil it down to a small set of principles.

Does that make sense? So his theory might be something like, "This person never uses words that are more than three syllables." Or, "This person tries to get alliteration into every paragraph," or something like that. He would squint at it and try to distill out what he thought was most important. Then using those principles, he would generate his own copy of the article, usually several weeks later. He would let the article drain out of his head. He would have taken notes on the specific, boring content of the article, like who said what, and what happened. Because that wasn't the part he cared about. He cared about the style of writing.

So he took notes on the boring stuff. He would come back to his outline, "Okay, here's the article that I have to write. I have to write that so and so said such and such." And then keeping these principles in mind, he would try to generate his own copy of the article. And then he would take that copy, and he would compare it back to the original. He would put the two of them side-by-side, and he would look at what he had written and what the person he admired had written, and he would find the places where he wasn't living up to the standard that he wanted to meet. And he would also find the places where he had done actually better, where he was a little bit smug about his own turn of phrase being even cooler than the other person's.

And then lather, rinse, repeat. He would take the data, distill the principle, attempt to act in accordance with the principle, double check. Maybe he refines his understanding of what the principles are, like, "Actually it wasn't this thing, it was more that thing," and looping through and through. Does this make sense as a concept? Yes? Okay. So you can't actually do this right now, probably, because you don't have all of this stuff. What I would like you to do, we're going to take another 45 seconds in silence. I would like you to think of somebody, or some skill, some place that you want to steal from, where you could actually employ this process.

So some blog author that you really would like to copy, or some athlete who had videos on YouTube that you could watch. I used to steal parkour technique in a fashion similar to this. You're looking for a place where someone has something that you think you could steal, model, mimic, and compare. To become a better person, according to you. Better at this skill that you want. Anybody uncertain about what I'm asking you to do? You're brainstorming a plan for the future, opening up an affordance for yourself.

Audience: Is it always with the same piece of material, or different material?

Both. Yes. Sometimes you'll repeat with the same exact piece, and sometimes you'll repeat with a different piece by the same author. And sometimes you'll repeat with a different author. The overarching thing is do what makes sense. You never want to go through motions just because you're supposed to go through the motions. Remember, do you know what you are doing, and why you are doing it?

If it's just because Steve said so, Steve better be real darn good, and you better be real sure that you understood exactly what Steve wanted. It's very easy to get lost in the weeds. Cool, 45 seconds, brainstorming. What do you want? What can you steal? Go.

Cool, raise your hand to the extent that you've found something. Neat. Raise your hand if you already... immediately, you're like, "I know. I know what I want." Once again, in pairs or triplets, just swap for a couple of minutes...

You guys can be louder than that. Come on, hum... there we go, humoring the sixth grade teacher in me, thank you. Quick double check. Do you know what you are doing, and why you are doing it? I want you to take 30 more seconds with your partner or triplet. And like, "Why is this thing worth learning?" What's the glorious, golden, juicy thing that is behind your desire to have this skill, or have this experience?

It doesn't have to be profound. There's no... I'm not being smug about, "Well, this one's altruistic, and that one's just self-congratulatory." But what's the... does it make your life more beautiful? Does it make you a more effective person? What is the reason that you want this thing? 30 more seconds each.

Neat, if I can get the mic ready... I just want a handful of people to share the thing that you want, and why it's good. Just raise your hand if you're willing to share in front of everybody. Yes.

Audience: My role model was Bill Clinton, and I would love to have his mastery of relationships. The ability to connect deeply and immediately with people in a variety of circumstances.

Okay, cool. Wait, you forgot to say why.

Audience: There's really two parts. One is that relationships are the secret to happiness, and great relationships can give you a very satisfying life. But also, being able to connect with people is a means to an end. Whatever you want to do, it's unlikely that you can do it alone. And so, if you're doing it with other people then you need to be great at relationships.

Anyone else?

Audience: I would like to be able to write and produce content effortlessly. And my two heroes, or people that I want to steal from in this regard were Tyler Cowen, and Scott Alexander who seem to be able to just magically write and produce extremely insightful and conversational blog posts in their sleep.

And what's good about the universe where you have this power?

Audience: It opens up a lot of professional and influence opportunities to build your personal reputation and brand in a way that scales. And also to communicate whatever message you want to communicate, to as broad of an audience as possible.

Nice. By the way, just in reference to Scott Alexander in particular. As my colleague was developing this techniquelet, the thing that he did it on to test the model was, How to Title Blog Posts as Well as Scott Alexander Does. He practiced creating... He came up with all these principles of how Scott Alexander makes his titles. Maybe one more? Yes?

Audience: Yeah. Mine was being able to develop questions that are very insightful, and my role model for that was Sam Harris, who in his podcast is able to generate really insightful questions that aren't answered by the authors, who have written books or newspapers. No, I'm sorry, newsletters are something else. Yes, his able to come up with these really good questions. And the reason why I'm interested is because it's important to be able to have good conversations with experts, and even in job interviews being able to ask questions that haven't been answered anywhere.

Cool, very good. A regular thumbs up. One thing I'll point out is that nobody, none of the three people that I chose had as part of their motivation wanting to feel good about themselves, or being able to be smug, or proud, or something like that. Which, it's possible that that motivation wasn't in the mix. But surely, for at least one person in the whole room, that motivation is in the mix.

And I don't want you to be unwilling to look at it. Good faith principle. I think your desire to be smug, and your desire to be awesome, and your desire to feel good about yourself are also things that should get a thumbs up. You don't necessarily let them steer. You don't necessarily give them the steering wheel. Then if you're optimizing for smugness, you'll probably do bad things. But if you have in you a desire for smugness, you can channel that and use it for good the same way that you can channel your desire for sugar into eating more fruit.

Cool. All right, moving ahead. Third step, how to invent your own rationality. So we have looking at yourself. We have looking at other people. And now we have, looking at the world. Squinting at the whole world, seeing, building models of how it works, figuring out your own understanding of what's going on in a situation.

And in particular, finding the places where it seems obviously broken and bad, and why is everybody doing it so wrong. It's almost incomprehensible. How could they be doing it so wrong? Nod if this resonates for you at all. Yes, this is where we've invented something like a third of our rationality techniques, is just one staff member who doesn't do it wrong, because of luck, because of experience, whatever. They're just like .. they're finally just like, "I can't stand it anymore. I'm going to make a technique, and I'm stop all of you from..." And it's great.

For instance, I'm not going to say this out loud because I don't know what the standards are about swearing. But probably my next essay is going to be titled something like this, because I just can't take it anymore. So, again, we're going to take just 30 to 60 seconds. Where is the world just obviously broken in particular?

Sometimes you already have this. You know exactly what's going on, and you could model it, and you could explain it to an eight-year-old, and the eight-year-old would learn from your wisdom and never get it wrong themselves. Sometimes it feels like this is impossible, like you'll never be able to crystallize out a model. It's just too large, too complex, you can't really get your fingers on all the gears.

I'm looking for the places in the middle. Places where you haven't really figured out what's going on, but maybe if you sat down for a Saturday afternoon you could. Does that make sense? Things that have been nagging you, bothering you, they occupy your shower thoughts. That sort of thing. Take 30 seconds, see if you can find such a thing in your own life, in your own experience of the world at large. Go.

Nice. Okay, I'd like you to take a moment, and imagine the universe in which I don't say anything else, and you have eighteen minutes just siting in this room in silence to invent your own rationality. To actually take, "Okay, Duncan's laid out some principles, and some threads that I can pull on." I want you to connect with the world where you in fact tug on those threads, and go forth and try to do something to tinker with your life on some level, or maybe on all the levels.

Can you imagine, in other words, can you imagine yourself taking these last 40 minutes and actually having a better life as a result? Take 10 seconds. If your hand is down here, this is, "I can't imagine it. That's dissonant, it doesn't make sense, it doesn't feel like how the world works." And if your hand is up here it's like, "Yes, I can see that, I'm excited about it. That feels real." Nope to yep. Cool.

Now I want you to imagine the world in which... let's see, it's June right now. Imagine it's September, and you've literally never again touched on anything that I talked about for the last 40 minutes. It's never come to mind, it's never been useful, it's never come in handy. You've done nothing with it. Imagine that world for a moment. How surprising is that world? Nope is like, "That's incoherent. I can't possibly imagine." This is like, "Yeah, that's the real world."

Okay. So in response to those dual imaginings, I have one thing to say, and I bet you already know what it is. Do you know what you are doing, and why you are doing it? Right. If there is useful content here, and you expect not to do anything with it, then like, "How about that?" Right. It's important that it be like this. It's important that it not be like, "I, fucking, again... I wasted my tools!" Because that's ignoring the parts of you that are leading you on your normal track. Your everyday autopilot is actually pretty darn good, right?

It's gotten you this far, and you don't want to just spit on that. Does that make sense? So not like, "Gosh, once again I didn't do what I intended to do." But more like, "I did..." Here I'm projecting, hopefully, "I did think there was useful stuff in Duncan's talk, but I can see myself never doing anything about it. What's up with that?" Right? That's the question that I want you to hold and ask.

And the way that I want you to hold and ask it is, do you know what you are doing, and why you are doing it? I want to pop up all the time for the rest of your life in annoying fashion. I want you to be walking down the street, and then little Duncan pops up on your shoulders like "Do you know what you're doing, and why you're doing it?" Next time you're eating the Oreos, next time you're road raging, next time you're getting in a fight with your manager, I'm going to pop up. I'm going to be there. It's going to be hideously annoying.

You guys are going to email me five years from now. You'll be like, "It's still happening." Do you know what you are doing, and why you are doing it? That's the key question. Makes sense? Cool. So all that being said, all of the how you should invent your own rationality, and how you should do this work yourself. By the way, if you don't want to do the work yourself, come to a CFAR workshop. They're super dope, and we will give you all the techniques.

But just in case. Just in case you never make it to a CFAR workshop, you now have the beginnings that you can follow up on should you choose to, and should you be able to figure out how. I'm give you one last lens. This is actually one of our techniques. TAPs, Trigger Action Planning.

Normally we spend over an hour teaching and working on this. So, again, I'm sort of just giving you the key, and you have to go further yourself. But hopefully this will be a useful tool for you to take away with you. If you want to learn more about TAPs, Trigger Action Planning, you want to look up implementation intentions. The reason we call them TAPs instead of implementation intentions, is when we call them implementation intentions everybody refused to learn them, and rated them at 3 out of 10.

When we started calling them TAPs, everybody loved them and glommed onto them. And it was like, "This is the best technique!" So we call them TAPs. Humans, go figure, right. But okay, here is the core of TAPS. You ready? I'm not very good at faking sneezes, but if I were to sneeze...

Audience: God bless you.

Right. So there's this immediate if/then that happens in your brain. Trigger/action, stimulus/response. When you hear "shave and a haircut" your brain is like, "I know the answer, the answer is two bits!" Right? When somebody sneezes your brain is like, "I know what I'm supposed to do here." And this has accumulated. Some of these TAPs, some of these Trigger Action ... This acronym is overloaded. This also stands for Trigger Action Patterns.

Some of these Trigger Action Patterns in your brain, some are there from biology, right, like the knee reflex. That's just there. That's a TAP that you didn't put there, your body just has that. Some of them are learned. Have you ever had the... let's see, I'm going to draw this real quick. Yes, ever had the experience of you're waiting at a stop light, and there's the turn left lane, and there's the go straight lane, and here's you in your car.

You're going straight at the stop light. And this little green arrow comes on, and your foot comes off the brake. Nod your head if you've had this experience. Because your brain has learned, a flash of green in your visual field when you're at a stop light, that means it's time to go. But uh-oh. Your TAPs can be more or less calibrated. They can be more or less appropriate.

Nod your head if you've ever gotten caught in the loop where somebody is like, "How are you?" And you're like, "I'm fine, and you?" And they're like, "I'm fine, and you?" Right? TAP, TAP, TAP. Your brain just chains together these if/then patterns into actually quite astonishingly complex behavior.

You can run on autopilot through the vast majority of your life, and in fact, you do. This goes back to where I was saying your autopilot's actually pretty darn good. It's super cool. Most of us who drive cars, drive cars almost entirely on autopilot in this way. Unless the traffic is really bad, or it's raining really hard or something, most of the time you're just doing it.

You can have a conversation with a person. You can be listening to the radio, and you're not even being an irresponsible driver. Your autopilot is actually competent to handle checking your mirrors before you change lanes. But there's this thing that can happen. My colleague, Anna, likes to tell story where she was driving down the highway at about 75 miles an hour. And there was a refrigerator in the lane in front of her. And she says she experienced "an awakening of sorts".

Where she had been on autopilot, but all of a sudden she was conscious. Like, "What was that? Was that... that can not possibly have been refrigerator." And there's this way in which, sort of like the Tesla autopilot assist. Like the self-driving cars that we currently have will be like, "Hey, driver, wake up, put your hands on the wheel." They'll call on the human at the appropriate moment.

Anna's autopilot summoned Anna, the conscious, verbal, explicit Anna. This make sense? It called her attention up. And this is one of the things that we want our autopilot to be able to do. There are two things that I want you to think about with your autopilot. One, where is miscalibrated. Where is your autopilot generating the wrong suggestions, the wrong actions. And two, when do you want your autopilot to turn itself off?

My colleague Val likes to refer to this as the summon sapiens spell. You can actually program into your autopilot, triggers for turning the autopilot off, and bringing conscious you online. In the case of Anna, it was accidental. It was just a refrigerator in the middle of the highway. But you can build up yourself some reminders. One of my favorite examples here, is they used Trigger Action Planning, under the name implementation intentions, as an intervention of ADHD middle schoolers, and it worked as well as medication, which is shocking and astonishing. This is actually real psychology, hundreds of studies, large effect sizes, and it passed the replication crisis.

This is still real, even after the replication crisis. What they did, imagine an ADHD middle schooler... this story that I'm now telling is somewhat made up and apocryphal. This is not literally what happened in the study. You should go read the study if you're interested. But this is sort of like the Sunday school version of the story. You've got a middle schooler who's just always losing track, they're just always off-task. And they're like, "I don't know. I don't know what happened," Like, "Okay, here's what we're going to do. Every time you notice that you're off track, every time your brain naturally comes to attention, I want you to review as much as you can of the last 30 seconds." That's it, just try to remember what was going on for the last 30 seconds.

And the first time you do this the middle schooler is like, "Okay. I notice I'm off track. I was staring out the window, and before that, I got nothing." And you're like, "That's okay, that's okay, good job. You got the first step. High five." And then they keep practicing this. And they're like, "Okay, I was passing a note to Carl, and before that I was looking out the window, and before that I dropped my pencil, and before that, I was paying attention. Yeah, I was paying attention."

You're like, "Okay, your pencil fell off the desk, and that's what distracted you." And then again, the next day is like, "Okay, well, I was tapping my foot under the desk, and then before that I was looking over at Suzie because she coughed, and before that, I was paying attention." And you start to build up this pattern of like, "Ah-ha, what's happening for this kid is noises pull them off task," right. When they hear a noise, they get distracted. And so, then you can build in a new set of Trigger Action Patterns. This is what a Trigger Action Plan is, is when you devise a Trigger Action Pattern. Either by replacing an action, or by building a wholesale new TAP, and you eventually get it embedded into your system.

Now this kid can be like, "Okay, here's what I'm going to do. When I hear a noise..." That catches their attention already, right, "When I hear a noise I will..." What do you think. They shouldn't just focus... they shouldn't be like, "When I hear a noise I will ignore it, and I will focus on the teacher," because that's fighting human nature, right?

But, "When I hear a noise, I will investigate its source, and once I know what it was, I will look back at the board." You can build up this TAP, you can build up this pattern, and this effectively, like, bam, ADHD cut in half. There's still ADHD, but at least this one problem is somewhat solved. They can come back to the board.

In particular, they're not fighting their own like, "I'm distracted." What they're doing is they're practicing the movement of returning their attention where they want it to be. Like the thing where people meditate, you're not supposed to fight yourself from thinking thoughts. You just think the thought, but then come back to your meditative state. You're practicing staying in the flow.

So, the shape of TAPS is just looking at what your autopilot is doing. Does it make sense? By looking at the triggers, looking at the actions. "When my significant other makes that face, I ramp up and start yelling. That's bad. When somebody cuts me off in traffic, I immediately narrativize that they're a bad person. When I have a package of Oreos in front of me, I put them in my mouth until they're all gone," right?

So this is a lens that you can use to look at your behavior, and then very carefully, very slowly, tinker. Now, a little bit of a word of warning. As you tinker, be careful and kind to yourself. CFAR used to recommend for people that wanted more exercise like, "Okay, you work on the fifth floor of a building. How about, let's try a TAP where when you get to work in the morning you take the stairs instead of the elevator."

And it turned out that the TAP that people were installing was, "When I get on the elevator I will feel guilty about not taking the stairs." This is bad. In fact, what you want to do is not use your TAPS as like a bludgeon against future you. What you want to do is have them be like a little dialogue box popping up like, "Hey, you had a goal of taking the stairs. Would you like to take the stairs?" That's what you're looking for.

You're looking for the autopilot to catch your attention, turn itself off, and bring conscious you into the frame. And then conscious you should make a choice. And if it turns out over the course of a week that your TAP does in fact fire. Every time you open the door to your building you're like, "Oh, yes, take the stairs." If it turns out that your TAP fires every day and you took the stairs zero times, that just means that you have a problem that's not solved by TAPs. Does that make sense? It means that you're like, "I thought that taking the stairs would be a good plan, but every single time the opportunity arose there was some resistance, or some problem." That means that you need goal factoring, which you can learn from a CFAR workshop.

So what I'd like to do with this last couple of minutes. First, I'd like everybody to take a minute, and I just want you to look at your own autopilot. Just try to see what TAPs you have. Do you have the tap of like, "I'm bored on Facebook," close Facebook, open new tab, open Facebook? So walk through your day, or look at life domains. What TAPs do you have surrounding exercise? What TAPs do you have when you talk to your friends? What TAPs do you have around eating?

I want to take 90 seconds. I just want you to look at the autopilot. See if you can notice these if/then patterns at all. Sometimes they're hard to spot. Go ahead. Nice. I'm actually going to hold questions for office hours. I have office hours in the co-working space from 1:00 to 1:30. And if you want to ask more about this stuff, or CFAR in general, or anything at all really, please feel free to come to those.

My last point that I want to make, you're going to walk out of this room, and everything's going to fall out of your head. You're going to drop right back into your normal autopilot. You're going to see people, you're going to go to the bathroom, "Where's my next talk?" And this is fine. This is fine. You don't have to be conscious all the time. Autopilots are really good. But there's one question that I want you to keep in mind, and I want you to keep hearing it in my annoying voice. I want it to keep happening forever, for the rest of your life. Will you say it with me? "Do you know what you're doing, and why you are doing it?"

If you keep hold of that question you will be improving your rationality all the time. Thanks everyone.