JOSH: Hello, and welcome to Clearer Thinking with Spencer Greenberg, the podcast about ideas that matter. I'm Josh Castle, the producer of the podcast. And I'm so glad you've joined us today. In this episode, Spencer speaks with Lynette Bye about the back-chaining approach to achieving goals, mitigating uncertainty when making life decisions, and developing a sustainable work-life balance.

SPENCER: Lynette, welcome. It's great to have you here.

LYNETTE: Thank you. It's great to be here with you.

SPENCER: So the first topic I want to talk to you about is the idea of back-chaining and how it relates to theory of change. Can you tell us a bit about that?

LYNETTE: Yeah. So back-chaining is a particular method that goes within theory of change in my mind. Back-chaining is when you take the end goal, say you want to solve AI safety problems, and you work backward, step by step for, Okay, in order to accomplish that goal, what do I need to do? What is the last step that would make it happen? And you do this repeatedly until you get back to things you can actually do right now?

SPENCER: Let's say your goal is like, I want to become a doctor, right? To back-chain, you might say, well, in order to become a doctor, I first have to graduate from medical school, right? And then you're like, oh, but in order to graduate from medical school, I need to, let's say, do a pre-med program or something like that. So that's kind of the back-chain process, you go all the way back to today. And you're like, okay, what do I have to do right now? So I can get their pre-med programs and get into med school and certainly, you know, become a doctor.

LYNETTE: Yep. And the back-chain is often going to be pretty tenuous, particularly if the goal is big and far off. So the theory of change, is this bigger picture of what is the system that you're within? And how can you use your knowledge of that, to make this back-chain work? So this is, for example, knowing that you can be a doctor is going to involve a residency after med school, and then licensure, you need to take tests before to get in all that kind of thing. And based on your knowledge, you're going to be able to make a better theory of change, the more you know, so often, the first steps when you're just starting out in the field are learning a lot. And that's before you can really make a fullback.

SPENCER: Okay. Can you elaborate a little bit on the theory of change? Sure, like, so? What does that mean exactly a theory of change.

LYNETTE: Yeah, so in my mind, it's the plan for how concretely you're going to accomplish your goal, really grounded in a deep understanding of the field that you're working in. This is your personal story and model for what needs to happen for that goal to be achieved. So if you're in book publishing, and you're trying to get a book out there, you know this is going to mean knowing, okay, I need to write a proposal and pitch it to the publisher, rather than just go and write the book and try and get it published. Then it's that kind of thing where you know, the field, you're able to use this because you have your own model. And that model is really necessary because, without it, you're just guessing or following the same steps as everyone else, which usually doesn't include enough of that nuance to get you to whatever particular goal you're going for because that's a very narrow path. I often hear of AI safety, as it's like, a general monolith that people like, oh, I'd like to get into safety. When I talk to people who are working on safety, there are a bunch of different fields. And like each organization has its own methodology and research agenda. Often, there are multiple research agendas that are very different going on within one organization. And which of these you care about, I think is important, is very highly dependent on what is your model of the problem that would arise if we made artificial intelligence, and it didn't go well.

SPENCER: So it's theory of change that is your mental model of how you get from point A to point B. So just bringing it back to the doctor example, for a moment. The standard path is like, oh, I need to, you know, go and get in, do a pre-med program. And then I need to go from there into med school, etc. But a theory of change might say, well, what actually makes someone succeed, to get into med school, you know, a lot of people try to get in and they don't get in, and what actually makes people succeed in med school. And then what actually allows people to go from med school into actually getting a good position as a doctor and so on. So it's kind of like digging into the details of, like, how that thing actually happens. And then think about that when you're actually formulating your plan.

LYNETTE: Yeah. So here, with that example, that might include knowing, okay, maybe letters of recommendation are really important for getting into med school. So while I'm doing the pre-med program, I'm going to cultivate relationships with some of the professors and do research with them so that they know me well enough that they can write a great letter of recommendation.

SPENCER: I'd like to contrast this idea of back-chaining from forward-chaining, where I think of forward-chaining as you're standing somewhere and you look around at what are your options, where you can move and you'd like to take one step forward, right? So this would be like, oh, if you're already enrolled in college, you're like, well, what classes are they currently offering that kind of point in the direction of becoming a doctor, right? But you'd have a biology class, you know, that kind of thing. And fortunately, forward-chaining is like you're trying to go start where you are right now and, like, build a path towards your long-term goal. Whereas backward-chaining is like you're conceptualizing your long-term goal, and you're trying to build a path backward to where you are now. Does that sound right to you?

LYNETTE: So yeah, that is how they're typically used. I have a little bit of a concern with that. Because when I see people trying to forward-chain, they're usually looking around. But they don't think, how does this connect step by step to the longer-term goal. So I think that even for forward-chaining, you want to at least have a vague idea of where this might go to connect to your goal, which is going to involve back-chaining. So basically, I think they work better when you're doing them together.

SPENCER: Right. It seems to me like they're both really powerful in their own way. Like, if you're just backward-chain, you may not take the best advantage of all the opportunities you have in front of you. But if you just forward-chain, you might take a super meandering path that doesn't actually move you that close to your long term goal, it's sort of the intersection of the two of them or really alternating the two of them regularly, that kind helps you build the optimal path. Would you say that's right?

LYNETTE: So I think my model is that you look around at the options and run these through a back-chain, where it's like, okay, would this connect to that goal, not just might this lead me in that direction. But once you know enough, is this an unnecessary step to reach that goal? If you don't know enough, then does this seem like a good option that will keep open the things that I could do later. And let me learn more is a really great way to go. Just gaining that knowledge and the information value is going to help you figure out what your next step will be. Basically, doing both looking around, running that through your own mental model, so that you're trying to take your best guess of what might be useful later and acting on that.

SPENCER: Lately, I've been enjoying using these kinds of embodied metaphors, where you imagine yourself in space, and you try to relate that to the thing you're talking about. And the reason I find them really powerful is it seems like the human brain is, like, extremely good about reasoning, spatial things like it's almost like what our brains were meant to do to a significant extent because we had to navigate the spatial world. And it's kind of similar to the idea that it's a lot easier to remember things, a lot of times if we visualize them, you know, it's like people who are professional memorizers and competitions, will associate like a visual image with each of the cards in a deck, right? And so then, if they get jack to five, that actually tells a story to them that they can visualize. It's not just a random sequence of cards. And that allows them to remember a really long string of cards. And similarly, it feels to me like we can hijack more of our brain to do interesting work if we can take this kind of an abstract concept and stick it into the world. And so where I'm going with this is that when I think about forward-chaining, backward-chaining, my like embodied metaphor for that is you're standing somewhere in space, and you're trying to get to the top of this mountain that's far away. And one thing you can do is you can, like, look right around where you're standing. And you can see, okay, well, I know I'm trying to get to this, like far off place over in that general direction, which path that like, I can literally see with my own eyes, right now, it seems to be like the best path because it kind of points in the right direction. And it doesn't seem particularly difficult to traverse, and so on. And so the forward-chaining is, you're like, okay, well, that seems like a good path to choose generally pointing the right direction, and you take it, whereas backward-chaining, you can't be on that far away mountain all but you could maybe like pull out a map or something like this, or ask people about it. And then by looking at your map or getting the information about this, like, faraway mountain, you could try to plot out a whole route from like that far mountain back to where you're going. And then that could inform the route that you take. Do you think that that metaphor works?

LYNETTE: Yeah, I love it. I think here, what I would be recommending is you look into the past, that are branching out from around you, see which of these seem good, and then compare them to the map and see if they lead you later in the right direction as far as you can tell.

SPENCER: Right. So you're hoping that the path that you choose now eventually links up with, like, your hypothetical path on your backward-chain, right? Yep. That you develop through talking to people through your map, and so on. And actually, this relates to an algorithm that's used in mathematical optimization, which is the idea of gradient ascent for going up or if you're going down, it's called gradient descent. But it's basically the idea of wherever you're standing right now, you say which one little step can I take that goes up the fastest, like, I'm trying to get this up a mountain, okay, I'll take one little step in whatever direction goes up the fastest, then I'll look around again and figure out now in my new position, what little step goes up the fastest. And you can prove that for many types of a mountain, so to speak, this algorithm will eventually get you to the top of the mountain. However, if there are lots of different mountains of different heights, and you really want to get to the top of the tallest mountain, this algorithm tends to get you stuck and will go optimal. Like you'll get stuck at one of the peaks, but not the highest peak. So that seems to me like a way of like kind of formalizing this idea of like forward-chaining is, it's kind of trying to say, which of my options available seem to move me up towards my goal the fastest, like, just picking a little step, I'm not taking a huge leap, I'm just taking a little step in that direction.

LYNETTE: Yep. The problem there is that in the mountain, you probably only have so many ways to go, wherein life there usually seems to be about 15 gazillions.

SPENCER: Right, exactly. Yeah, there are way too many. But not only that, some of the best ones are the ones we haven't even thought of. So, this reminds me of this idea of narrow framing, where it's really common, that when people are thinking about a decision, they frame it in terms of like, do I do A? Or do I do B? Like, do I quit my job? Or do I keep my job? Right? And it's this kind of thing. But a lot of times, there's some other option that's maybe a hybrid, or is kind of out-of-the-box thinking or whatever that actually might be preferable than that either, like the two standard options that you're likely to think of.

LYNETTE: Yeah, at least for example here of career decisions, writing up your decision, and then sharing it with some people you trust to get feedback, can be an interesting way to generate new ideas, often just sitting down and generating a bunch of self will do a similar thing. But for getting you out of the limits of your mind, there's this interesting thing in psychology, where I always take psychology things with a small grain of salt in case it hasn't been replicated since I was an undergrad. But there is this idea that once you were thinking of one idea, your brain, because there are so many overlapped concepts stored in your neurons, it actually suppresses the nearby ideas. So when you're brainstorming, and you think of one idea, and that leads you down a chain of here are related things, but there are other ideas that your brain is actually stopping you from thinking to keep on this one track. So this is actually making it more difficult to brainstorm widely.

SPENCER: That's really interesting. I feel like I've experienced that just anecdotally many times where I'm trying to think of someone's name, and like, I think of it like almost the right name. And it just seems to block. I think there's something to that. So we actually did a study on narrow framing. It was pretty funny actually, the way it worked is we had people write down a big life decision they were going to make and generate possible options for what they could do. And then half the people have chosen at random, where it was suggested to them that they might want to come up with more options. And you know, we've explained that sometimes people don't join options. Actually, half of that group did it. They were pretty obstinate and refused to do it. Because I guess it just seemed boring, and like why bother, right? And the other group, we actually said, Okay, we're just not letting you proceed and study until you generate at least one more option that you haven't thought of yet. And in fact, in that second group, we forced them to generate at least one more option, about 20% of them at the end of the study said that they actually thought that was the option they were going to take. It was kind of remarkable. I think we can be pretty confident that the first group would not have generated them because we literally gave them the opportunity to and they didn't bother. This is a small study, you know, all the normal caveats apply. But I do think that's really true. I have been there many times where I've been talking to a friend, and they've framed the decision in terms of like A or B, and I'm able to come in and say, oh, you know what, maybe they're C and D and like, maybe they're at least worth considering. And I think I did the same thing. You need something to break you out of whatever framing you start with.

LYNETTE: Yeah. When I'm talking with people who are usually earlier in their career, and they're trying to think of what things they might want to do, I'll encourage them to go with and think up between 15 and 100 options. That dazzles them. But the idea is that most of these aren't going to be good, they'll overlap with each other or be wildly impossible. But you've just to come up with them. And you're going to think of a lot of things that you wouldn't have if you were just making a list of four or five.

SPENCER: And so the will of those be like fully fleshed out, like, well, I could be a dentist or you know, I mean, like, or what kind of ideas are that you're having them generate.

LYNETTE: So usually, it's a combination of things, like it's, oh, maybe I could do one of these six things during summer off. Or I could do one of these four internships, or I could do these 17 projects or informational interviews to explore these different paths. And so it'll be an overlap of fully fleshed out ideas of tests that they could run on a bigger idea, but that itself is smaller, or just like different versions of something I could do like, last time I did something like this, I think it was like, well, maybe I could write a blog, or do a workshop or make a course. And then they're like, that was the first of like 20 things that I brainstormed that were adjacent to coaching that maybe would be useful.

SPENCER: So I think when some people hear about this, they might say, well, when that'd be overwhelming to, like, have so many options, and how would you even reduce them down? So then, what do you advise them to do once they have all these options?

LYNETTE: I mean, realistically, by the time they have the full list, at least half of them are just going to be okay, that's implausible or clearly worse than these other ones. But at that point, they have enough that they can think through all these other ones, and see if there's any that they've generated that does seem really exciting. Like, I've had people say that they generated a list, but then they talked with someone who was like, oh, wait, why aren't you considering this very obvious other things, that would be great. And that became their top choice. And just the fact that they had only made a small list meant that they weren't considering these other things that were slightly outside of the, here's the option right in front of my space.

SPENCER: Yeah, I think that's a really nice idea. And I think in general, people feel like creativity has to be this, like, magical thing that just happens to you. Like, I'm just walking around or in the shower, and suddenly, like my idea for my creation, pops into my head, or my idea for my startup or whatever. And I almost completely disagree with that, as someone who will just kind of sit and is like, okay, force myself to come up with an idea, like, next five minutes. And it doesn't always work, but like, a lot of times it does. And so I tried to develop tricks for, like, getting myself to think of ideas, but I really believe that you can actually just sit down with a timer and just generate ideas. And if it doesn't work at that time, you can come back an hour later, or 10 hours later and try it again, and often it will work. One thing I find really useful with that kind of brainstorming is actually to give myself different prompts. So instead of, just saying, okay, think of an idea for a career I could take,” right? I like to try to give some prompt like, okay, what about a career related to artificial intelligence? Okay, what about a career related to psychology? Okay, what about a career related to business, you know, and actually, having that seed can actually make the brainstorming much more efficient. It's kind of fascinating how that happens. Another way to do it, which can be even more powerful, is if you can subdivide the space of all plausible options up, like cutting it into mutually exclusive collectively exhaustive units. So for example, like, you can find binary splits like, well, it either is going to be a career in let's say, for-profit or nonprofit, right? Let's say it's, you know, we've one split, okay, we've subdivided the space, okay? Now, within for profit, it's either going to be earning to give money away, or it's going to be a company, that thing is high-impact or whatever, right? Whatever your personal goals are, are going to determine the splits to some extent. But once you've kind of split up the space into a bunch of units. Now you can say, okay, within the space of exactly this type of thing. Let me try to brainstorm and then you can move to the next empty bucket, and then try to brainstorm on that, and so on.

LYNETTE: Yeah, so there's a similar concept. Richard Hamming, a pioneer computer scientist, had this method where he set aside every Friday afternoon as his great thoughts time. And during this time, he would give himself prompts for things like, what is going to be the future of computers, or what is fundamental in my field that I don't understand.

SPENCER: That's so cool.

LYNETTE: And he would use these to try and really deeply understand and think about the big picture. And then all of that would inform what he did the rest of the time. But I really liked this because that concise setting aside of times seems like the best way to facilitate this kind of thinking on a regular basis.

[promo]

SPENCER: This trick that I've been using a lot lately, which actually Twitter caused me to do inadvertently, is, which is it's rare for people to talk about, like thinking benefits they've gotten from Twitter, but I really feel like I have, is basically I want to tweet things are valuable to people. And so I'll pick a topic like habits or decision making or the sunk cost fallacy or whatever. And then I'll try to say to myself, “What's the most useful thing I can say about this topic of 280 characters?” which is really, really hard to think about. But then once I've kind of formulated, something that I think is really the core of this idea, something valuable that I can squeeze into that few characters. I feel like I've encapsulated something important. And I can put it out on Twitter, and hopefully if people find it valuable, but the remarkable thing I found about this is it actually helps me understand the topic better because I'm forced to cut through all the noise and all this complexity and just get to like the core. And the really interesting thing to me is that once I do that, I find it's actually way easier to then go write an essay about that topic, because I already have kind of the meat of the essay, the core of it, and now you're just kind of building out the stuff around it. And so I've been doing this process where I call it tweeting the core, I first tweet the core of an idea. Then two things happen. One, I already have the core of an essay. But also I can see, does it resonate with people today who find it valuable. And if they do, then I'm going to say, okay, now I can turn the core into an essay. And then I can post that elsewhere. So I've just really loved that feedback loop, personally.

LYNETTE: I really like this type of method, where you have an idea. And you're starting with the smallest unit that you can get feedback on, just because there's so much going on, particularly when you don't know what's going to be useful, such as in research or writing. And you can just get that little bit of feedback. You've got the feedback from the audience, and you've also gotten the feedback from your thinking about it. Does this seem useful? Does it seem exciting to you?

SPENCER: Absolutely. I mean, it's so powerful, having tight feedback loops like that, for me, it's even, it's even crazier a feedback loop. Because usually then, when I am writing an essay, my next step is to post on Facebook. And then what happens is, a bunch of people will go critique it, or they'll suggest ways to improve it, or they'll riff on it. And then I have a whole comment thread or people like commenting on it. So then I take their ideas, and the ones I think are good critiques are good, like ways of building on it, I'll say, oh, thank you, that's awesome. I'm gonna add that to the essay, I'll go improve the essay based on their feedback, then I'll turn that final thing into a polished version that based on all the critiques that then will go on my website, it's kind of this multi-level chain, I think it's just wonderful to try to find, like tight feedback loops like that in your life, do you have any that you use?

LYNETTE: So I do know a number of things, kind of along this line. With writing, I've started doing user testing interviews, which is I'll take whatever the concept is. And then, I will sit down with someone and try to help them use it and just make it optimized for being as useful as possible to them. And in the process, I'll find out what parts they don't understand, where do they get stuck? What does the existing tool not help them do? And by the end of it, I can make a much better tool for whoever else after I publish it.

SPENCER: That's really cool. So it sounds like you're doing like a process that's usually done with products, but you're doing them for your essays, with ideas, to try to understand, like, how is this idea helpful? How is it not helpful? How can I explain it better? You know, how can I relate to people's experiences, and so on?

LYNETTE: Yep.

SPENCER: That's really wonderful. I actually did something like that when I was doing a TEDx talk. What I did, so we built a platform called Positly that she recruits people for studies. So it's really easy to get people for your studies. And what I did is, I recorded a really quick and dirty version of like a very rough draft of my talk, I then threw it on the Positly, had, I don't know, 20, or 30, people watch it, I paid them in a small amount of money, they watched it, and I had them fill out a form saying what they liked and didn't like and critiquing it, then I took their feedback, I improved my talk, I then did another rough cut, basically, with the improved version, put it back on the Posilty again, you know, got another 20 or 30 people to critique it. And then by the time I'm going to get my real TEDx talk, you know, I've had like maybe 50, or 60, people watch earlier versions, critique it. And so it's like, I already know what complaints I'm going to get, you know, any complaints that are left are ones that I just either have no clue how to fix, or I'm just willing to accept. But other than that, you know, like, there's no real surprises.

LYNETTE: So these loops that we're talking about are things that are pretty similar to what's done in product development. But I think that there are similar things that you can do to further out decisions, like making a career decision, or figuring out what your research topic should be, which are both big important decisions that don't have a clear path, and often depend on you finding ways to do stuff that other people aren't doing or can't do. So in these often, it's more things like, you do a little bit of research, and then write that up. And share it with a couple of researchers, you know, or just try and learn about an area and write a summary of it, where you're getting that little bit of feedback loop from trying to explain it to someone in this case.

SPENCER: Yeah, absolutely. I mean, often I think about there being three groups of people, you ideally want to get feedback from. The first is, like, your end-user. And we already kind of talked about that, like, oh, your audience for your talk, or your audience for your blog posts, or what have you. The second group is people with expertise. So like they have deep knowledge and whatever the subject is. So for example, if you're trying to make a plan for how to get into med school, you should probably talk to some people that went down that path, right? Maybe some people have succeeded, and ideally also some people that failed, if you can, because they often have useful insights, as well. You know, if you have a selection bias, you only talk to people who succeeded, you may actually miss out on important information. The third type is people that you just really respect their thinking or decision-making, or you think they're wise. So they may not have domain expertise, but they may be able to help you think about how to make that decision. Or ideally, also, they would give you people that understand you pretty well like, you know, the friends of yours or family members of yours whose decision making most respect, and they can kind of guide the process even if they don't have the expertise.

LYNETTE: Yeah. And when you're talking with the people who have that expertise, I found that you can find other things written about this that often you don't just want to ask them for advice. If you just ask them for advice, what you're gonna get is kind of filtered through their perception of, oh, maybe here's what other people need. Where if you instead ask them, what's your story? How did you get where you are? You're gonna come up with more detail of them, and maybe detail that you wouldn't have gotten just asking for advice. You'll get the, like, little tips and tricks that they've developed and are just kind of in the back of their mind, but turned out to be really useful to a new person who's trying to master this.

SPENCER: I think that's good. But I think it's also important to ask them things like, what do you wish you'd known when you were going through this process? Or whatever mistakes people commonly make? Another question that I think can be really valuable is to ask, if you were me, how would you think about this process? Or what do you think about this decision? Or how would you think about, you know, getting to this goal?

LYNETTE: I think in my mind, the ideal would be to do some of that. And some of us know about their story. I've heard of one method where you do both of these. But then you take that and make your best plan and ask them, “Is this a good plan? Do you have ways that I could make it better?” which is giving you even more concrete feedback here's the actual plan that I'm presenting to them getting feedback on. And it's also in a way, acknowledging and thanking them for the advice that they'd already given by saying that you're trying to put it into use?

SPENCER: Yeah, that seems really, really useful. That actually reminds me also, there's a key thing here, which is how do you actually get useful feedback because I think people really underestimate the extent to which most people don't want to give you negative feedback. So if they think you'd be amazing in that career path, like, they'll probably say that, but let's say they actually think you're gonna fail, they probably are not gonna say that, right? It's just, it's awkward, and most people don't want to make you feel bad. And your friends, especially, don't want to make you feel bad. So I've developed a number of tricks that I find useful for this kind of thing. One of them is to flip the conversation, where instead of them thinking, “Oh, I'm being a good friend,” or “I'm being supported by telling them they're great, and like, they're gonna succeed, and everything's awesome, and that they're doing everything right,” instead of flipping to say, “I really want your help to figure out what's wrong with my plan because I really want to succeed and like, the way you can help me is by helping me not fail.” So suddenly, that flips it to like, oh, okay, I've got to actually point out what's wrong with this, because the way I help them with a good friend is by actually finding the flaws with them. So that benefit is really useful. And when I do things at a distance, like let's say a survey, another trick I use is I say, right before the survey questions start, we want your brutally honest opinion because we want to know really what you truly think about this. I think that also kicks people into a mode of okay, I'm just I'm not being nice, like, or I'm being nice by being honest. Right?

LYNETTE: So I like the framing of the first one in particular because it is just telling them, frankly, that this is what I appreciate. I often just ask, “Okay, what's most likely to fail about this plan?” Which can get some interesting responses.

SPENCER: Yeah, that's a really good one, too. I'm just gonna directly be like, well, you know, that comes up in customer interviews, too. I'll ask potential customers awesome things like, suppose you don't use this product, what do you think the most likely reasons are going to be? Because if you say, would you use this product? Well, I'd be like, sure. Yeah. It's like, okay, suppose you don't, then why not? And I think people actually are pretty good at projecting the future and saying, Well, maybe it's a little expensive. And also, I would probably try out your three competitors, and so on.

LYNETTE: So another thing that I found is that asking people questions, it gives you what they think. But often I get a lot more information by having them use something and then just seeing how they interact with it. So when I give someone a draft I'm working on and say, does this make sense? I will usually get some version of Yes. Even if I'm trying to get negative feedback. But if I have them, read it, and then tell me, okay, here's what I would do based on this tool, I find that they're just doing something completely different than what I had in mind, because of some subtle difference in how they interpreted when I wrote, but that right there just tells me, okay, here's what I need to fix, or the caveat I need to explain so that the next person will understand it.

SPENCER: I think that's really good, too. And that reminds me of a technique you can use with products, which is kind of related, which is, just watch someone use your product who's never seen it before. And say to them, just speak out loud, everything that runs through your mind, you know, just kind of verbalize your mental content. And you just don't help the person, right? You just watch them fail. And then, of course, if you know, if it gets ridiculous and, like, not useful, the more you're going to help them but generally speaking, you just watch them fail and you take detailed notes about the way they fail. And anything they say it seems like they're confused or frustrated, or what they like, and so on.

LYNETTE: That's basically what I do with some of my tools. I do think there's a really important mindset here, which is something like the client is never wrong, but the person using it, assume that this is just the normal person and everyone will get a similar thing wrong and how would you improve the product to deal with that, rather than ever assuming that it's just this person.

SPENCER: Right. Although, I think you also have to balance that with the fact that people are idiosyncratic, right? So there's a balancing act there were, on the one hand, you really have to assume that if someone fails using your thing, you should take that very seriously. But you should also keep in mind that it could be something to do with that person at times. Just to give an example of this one time, I was really demoralized because I spent a ton of time making something. And I showed it to someone whose opinion I respected. And they also seem to me like the target audience for this thing. And they really thought it was crap. And they were like, I don't see why anyone wants to use us. And it was a really big bummer. And then I released it. And it was literally the most popular thing I ever made at that time. I think over a quarter of a million people ended up completing the usage of it. And I always had that back of my mind. It's like, there are these two biases on either side, right? Maybe the more common one is just ignoring negative feedback, that probably is the more common one. And you should take negative feedback extremely seriously. But there is a bias in that aside, it was just getting demoralized or over-emphasizing one data point. And I think actually, there's a really simple solution to that, which is just don't just have one person critique, right? Have like three or four or five. Five is a common number of people to talk about if you really want to get good qualitative feedback, at least a starting point. And then once you've done five, if you're like, you know what, I still feel like there's a bunch of unanswered questions, and maybe you do another five or something like that.

LYNETTE: Yeah, I agree with that. I think I am thinking more of where they are getting confused or stuck. And those are the things that I need to fix. Even though it often feels like I tried to directly address that, it was just clearly something that broke down there. So assuming that it's always me, helps me get in a better frame of mind to make sure that I'm capitalizing on this information.

SPENCER: It's kind of incredible how hard it is to forget what we know. If you wrote a blog post, or you made a user interface, it's, like, almost impossible to see it with fresh eyes. And so when someone makes some mistake, like, how could you make that mistake, it was so obvious? And yet, it's so not obvious, because you just come in with so much information that they can't possibly know. And actually, this idea of learning to see your own thing with fresh eyes, I just think it's just an incredibly powerful skill that I've never really heard people talk about in this context, at least that I recall. But it basically is like a superpower. Because if you make a thing, and then you can get into the mind of your user, and like, see it from the perspective of someone who's never seen it before, you can actually do that first feedback loops without a person involved, which is so much faster, such a faster iteration. So like the first five iterations are just using it with fresh eyes again, and again, finding problems with it making it better, then you can move on to actually involving real people. But now, you've already done these super-fast feedback loops where you didn't even need a real person.

LYNETTE: Yep. And this is called the curse of knowledge, the inability to get past what you know, to see it from another person's perspective.

SPENCER: Yeah, absolutely. And I think just more generally, our ability to simulate other minds just has so many benefits, like whether it's simulating a friend, my friend enjoyed that you know, or simulate your romantic partner or you know, but simulate your users. If you're trying to make anything that other people are gonna use, that's just, an incredibly valuable skill. And it's actually, I think, one reason why it's so useful, doing lots of customer interviews, like, often with the CEOs I work with, all suggest to them that they do two interviews a week. And part of the reason for that is because there are always important questions to answer and the interviews are supposed to help get them. But another part of the reason for that is that it actually just helps them build a mental model of the customer. After doing enough of these interviews, they can kind of simulate the customer and say, well, what would Sally have said? What would Joe have said, you know, about this new product or feature idea, whatever?

LYNETTE: Yeah, this is definitely more scattered and less focused. But I find that themes will come up commonly enough in coaching calls, that in one sense, I'm just doing customer interviews all the time for what are the biggest issues and what kinds of things help people get past them?

SPENCER: Yeah, that's fascinating. So why don't you talk a little bit about the services you offer? And then I'd love to ask you about what are the themes that you keep coming up with again, and again, in your coaching?

LYNETTE: Yeah. So I do productivity coaching. I'm starting to refer to it as productivity and career coaching because I've discovered that when I say productivity, I mean, everything involved in accomplishing your big goals. And when some people hear productivity, they think of to-do apps. So it's a much broader scope of things. A lot of it is prioritization, thinking through big plans, setting goals, and then, building up all the habits that people need to support that ongoingly.

SPENCER: Like a full-stack coach.

LYNETTE: Basically, focused around career stuff, most of it is one-on-one coaching, and then I do the writing and stuff on the side based on the things that I've learned that are coming up. So I did a piece on prioritization a little bit ago, that was just, I put together all these examples. And each of these examples was based on at least two conversations, often more than that, that had followed similar patterns as like, here's the problem and here are the kinds of solutions that were helpful, which in this case was how can they frame things in order to help them prioritize the right things to do.

SPENCER: Do you feel comfortable talking about just some of the general patterns that come up when people are thinking about their careers? Like, what are the common issues that people struggle with?

LYNETTE: So one of the biggest ones is basically not having the information that they need to make, to feel confident that they can make a decision. So this is someone who's thinking about making a career shift, but they don't know if they have the skills to be good enough in the career they're trying to go into to make it worthwhile. Or someone who's young in their career. They recently graduated from college, and they're considering a couple of different career paths, and they don't know how to decide between them. I'm working on a post related to that now, which is this whole muddle around combining some of the stuff we talked about earlier with a theory of change, with how do you get your best guests, and then narrowing down your key uncertainties around this and go and test them out. One, as I've talked with more people, this kind of iterated winding path toward what you pay for and ultimately did and found to be useful seems to come up quite a bit. Like, I think there's a decent chance right now that if I asked you how you came to host a podcast, you would have kind of a winding path that would lead me here. I'd be curious if that's true if you want to tell your accounts of how you came to do that.

SPENCER: Yeah. So basically, I've been thinking about doing a podcast for a while when I realized that there started to be successful intellectual podcasts, which made me believe that you know what, there might be a market for talking about ideas in an analytical way, which is something that I love to do. And I think that I know a lot of people who enjoy doing that as well. I was seriously considering it. But then I was also thinking, Okay, I don't know how to edit a podcast. And that's, you know, that part of it is not playing to my strengths. I don't have any audio editing skills, etc. And it just so happened, one of my work colleagues was like, you know, you should do a podcast, and I'd be happy to edit it.. So I was like, ah, bam, okay, that's perfect. That's how me and Josh Castle, who's the producer editor teamed up because I was already thinking about it. And he suggested an offer to be better. So yeah, so then I was like, okay, let's do this.

[promo]:

SPENCER: And then, as soon as I started, with the first experimental episodes, so I did about three. And then I recruited about 60, or 70 people to listen to the three first three podcast episodes, and then fill out, like, a feedback form critiquing it. And I wasn't really sure whether people were gonna like them that, you know, I've never been a host before, but we got quite positive feedback, or it was better feedback than I expected. There were definitely some flaws with it. But there were also things that seemed like I could get better at. And so that really gave me momentum and say, okay, I think I have something here that is a viable idea. And it helped me hone the idea, getting all that feedback from those people. And then I just started going wild. And you know, actually having so many episodes pre-recorded is really great, because it means that we can really launch with our some really strong episodes so that people really understand like, what is this podcast about? It's, better for him, not like me flailing around and having to figure out what the podcast is for him.

LYNETTE: Yeah, so right there, you had an idea. And it was floating around probably competing with many other ideas of things you can do. That coworker prompted you to like, “Okay, let's try this.” So you did a test, liked it, got some good feedback, and decided to go for it. And then you did a bunch more. So that is a whole iterative process. I'm guessing if we traced it back even further, there would be a winding path to lead you'd even have the idea of a podcast or wanting to do intellectual work, to begin with. But yeah, that's the kind of iterative process that I'm talking about, of how can someone take an idea, and for something that isn't a traditional product always, and go and validate it so that they have enough of a reduction in their uncertainty that they can go forward and make decisions that are at least probabilistically more likely to lead them to the outcomes and the impact that they want to have.

SPENCER: So when you're talking about iteration is the idea that you take one action that helps reduce your uncertainty about the plan or about the project? And then once you've done that, you take another action, maybe reduces uncertainty along with the different access or different part of the plan, something like that, or do you mean something different by that?

LYNETTE: Yeah. So there's a couple of different ways you can go. That's definitely one of them is, where you have a couple of different modes of uncertainty. And you kind of go through one by one and reduce it. This is a case where if somebody has five different research ideas, for example, and they all seem equally good, maybe they want to do a couple hours of research on each and see if after that they have a different ranking of what seems most likely to be useful. Another one is like a proportional investment where once you have the idea of, oh, podcasts would be interesting, you make a couple of episodes. And then when you have that validated, you make a bigger investment. And you kind of keep this cycle going. So that the more validation you have that this is useful, that you're on a good track, the bigger the investments you're willing to make into it.

SPENCER: So you kind of ramp up and gather evidence about whether you should further ramp up as you're wrapping up.

LYNETTE: Yeah, the idea there is that you're not going to wind up with 60 podcast episodes, before you ever realized what you were doing wrong, and discovered that because of that, nobody liked him or something like that.

SPENCER: It reminds me of what I think is both good and bad about the lean startup methodology. Like I think the lean startup methodology gets so much right about building products. For example, it says, you've got to get data, you've got to get information right away, you want to create it, create something and put it out there and get feedback, you know, just try things. And I think all that can be really, really good advice and combat, it's a really common problem where people spend way too long on something that really nobody's gonna ultimately want or is just kind of going in the wrong direction. And a huge waste of time and, you know, increases people's chances of failure a lot. But on the flip side, what I don't like about the lean startup is I think, it seems to sometimes suggest that the way you get information is like building the first version and like releasing it. Whereas, for example, I mean, I could be wrong about this, but I actually think my podcast is much better off because I didn't release it immediately. I didn't go create three episodes, and then release it and say, “Now I have a podcast,” right? Instead, I released three episodes, and I had 60 or so people listen to them and critique them. That was my evidence loop. In other words, I think there are a lot of ways to get evidence, I think it's incredibly important to start getting that evidence data like coming on, as soon as possible. But it doesn't mean putting out in the world. And I believe I could be wrong, I think my pockets are going a lot better because we're not going to release the first like the super rough cut episode, like episode number one, you know?

LYNETTE: Yeah. I don't think that that's in contrast with lean methodology, at least not the version of it that I've studied. It's very much about creating these tight feedback loops. But doing a batch case with a small group, before you put it out, it's still a totally fine way of getting evidence.

SPENCER: Maybe, I'm refining the way a lot of people think about it, which is like, you know, build the first version of your product, and like just put it out in the world, you should be embarrassed about it. Another related thing that, like, comes up, I think, where I feel like the Lean Startup method can be misleading is, it seems to de-emphasize thinking, and instead emphasize doing and I think for a lot of people that's really useful to emphasize doing over thinking. But actually, I think thinking is incredibly powerful. And like actually doing exercise like the one you talked about, where you actually think about 15 or 20 plans for your life that can actually be super powerful and can avoid a lot of problems that would occur if you just started like immediately trying to execute things.

LYNETTE: So when I'm thinking about this applied to personal life, I started with defining the decision, which includes things like variable change, figuring out what are your key uncertainties, the crucial things that will change your decision based on the answers to them. And then based on that you design tests, and a test could be as simple as I write a blog post and see if I enjoy writing it, I write there, it's giving me information that will inform whether this is something I should keep doing.

SPENCER: I had someone talk to me about whether they should go into AI safety research. And we were talking about the pros and cons. And I was like, okay, so have you ever tried doing SAP research? And they're like, What do you mean? I'm like, well, you know, you could just try it, like, at least for a few hours, you know, maybe a few days even. And it was interesting because it hadn't even occurred to them. It's like they thought, well, I can't do it. I see through research until I'm part of a team that does as if something it's like, well, but you actually could just try it. You know, like, I think a lot of times we could try things, maybe, you know, maybe we can't try being a surgeon, okay, that's unrealistic. But maybe we could at least, like, find out what a surgeon does all day. Maybe we could even get someone to let us, like, go to the hospital to watch a surgeon, you know, that doesn't seem completely impossible.

LYNETTE: Exactly. That is exactly the kind of experiment that I'm encouraging people to do.

SPENCER: Right. So what kind of self-experiments do you talk to people about?

LYNETTE: So getting feedback, looking at successful people, and trying to copy what they do? This is particularly when you have an idea of what you want to be accomplishing. But you don't know the path there. They're usually there's just a lot of options of things you could try. So finding out what people did that eventually led to them getting there is a good place to start looking for options.

SPENCER: I think that's good. But I just want to come on that for a second. Because it seems to me that if you ask someone a question like how did you get to where you are today, or like, what was your path that oftentimes people seem to come up with, like, explanations for why what they did make sense in retrospect, but they don't necessarily tell the true story of like, actually the causes that led them there. And I don't think that they're being deceptive or anything. I think it's, just like, there's a way of answering that question. It's like, well, why did I do this? Well, it's good to do that because of this thing. But at the time, that may not really reflect the actual decision-making process. What do you think about that?

LYNETTE: I think that's true. It's a delicate balance, sometimes it is still useful. The other thing that I do to get around this, is asking them step-by-step, walk me through it. So basically, tell me the story of what they did. You're still getting it filtered through their memory, what stood out to them, often the path where they are seems much more straightforward than hindsight, they don't remember all the things that didn't work out. But it's still giving you more context, at least on how did they get here, what were the things that they were doing at the time, which, if you're really getting into it, and ask them like, step by step to lay it out, you'll usually get more information.

[promo]

SPENCER: Sometimes when people ask me, you know, Spencer, why don't you get a Math Ph.D.? I respond, Okay, do you want to know, like, the actual series of causes that led me to go out this year? Are you asking what's good about getting a Math Ph.D.? Or, you know, because actually, they could mean, they can have a lot of different intentions for asking that question. Like, what are the good things about getting a Ph.D.? Or, you know, do you think it's justifiable to get a Ph.D.? Or how did you actually end up in this place, you know, casually, etc? These questions are often worth disentangling, depending on what the person is actually interested in.

LYNETTE: Sounds right.

SPENCER: Any other kinds of self-experiments that you recommend for people?

LYNETTE: Yeah, so definitely, as you were mentioning earlier, just try something yourself. Often, a small side project can get a lot of feedback on how much you enjoy it, and how good a fit you are. For some things, such as AI safety, there are a lot of tools out there, like open areas, spinning up a course that people can try out, dip their toes in the water, and see if they should try applying to places. So I'm adapting this from business literature, critical path analysis, where the idea of this is that you lay out as best you can tell the different steps that you could do. And then you structure them so that you're doing the ones first that most quickly reduce your uncertainty. So this is anything that might fail and say, okay, I shouldn't go down this path, you try and test that. And anything that would say, here's what you should do next, or change the other actions you'll take, you try and do that. When I was trying to start the blog, I had an idea that I might like to write, but I've never really done it. So I started with a small experiment of just posting on Facebook each day. And I explicitly limited this to what I could write in 15 minutes. And one, this gave me some feedback on, did people engage with the stuff I put out? but it gave me a lot of feedback on both the logical, just like, do I think this is useful? Do I think I have anything useful to say? And the emotional, I think a lot of times with career decision in particular, what's going on isn't logical, as much as emotional, like, “Can I do this? Would I be happy here? Is this going to be risky?” And sorting through all of those is as much a part of the test that you need to run as “Can I code?” So starting with that, and then after that, I took one of the things that I'd done this little post on, made one longer post, and put that out. And this was the proportional thing where I did a little bit more. And then the quarter after that, I decided and made a commitment to write one post per month. And I liked that. So I kept doing it the next quarter, and then the quarter after that. I've been running a bunch of other tests during the first two quarters. And then the third quarter, I was like, okay, I'm just gonna focus on writing. And so that's what I've been doing this past quarter, is my main experiment. And that's like, kind of an iterative process that I'm currently doing, based on this idea of first doing the things that will give me the most information, which like running a bunch of experiments and deciding which one I wanted to make a bigger investment in.

SPENCER: So just to give the example of why this idea of critical path analysis and kind of focusing on the biggest points of risk or uncertainty makes so much sense, like imagine that you're trying to manufacture a new type of item, right? And let's suppose that most of the steps in the manufacturing process are well-known steps, you know, they're used for manufacturing many different types of items. But there's one step that is new, that has never been done before. If it's needed for manufacturing, this type of item you're trying to create, it could be very easy to just focus all your initial efforts on getting in place all the standard manufacturing steps because, well, they're understood, they're relatively easy. There are not a lot of barriers. But if you do that, the problem is that it pushes this big point of risk way off into the future. And then if you can't actually overcome that risk, when you get to it, that part in the process, that's not well understood, then your whole thing fails, and you've invested all this time and money. So instead, you want to move that part that's uncertain, the one piece of manufacturing that is not well understood to the beginning of your process, even though it might be, like, stressful and feel more difficult and less actionable. Because then if it fails, well, okay, you barely are investing time, that's not a big deal. And if it succeeds, well, then you basically now can have much higher confidence that the whole plan is going to succeed. And you can kind of move forward.

LYNETTE: Exactly. Or for an example, that may be more relatable to many people. Often when someone's in college, thinking about their careers, they have tried out something and they like it, and they'll do a lot more of that. I definitely did this where I love psychology. So I took a ton of psychology classes, which I enjoy taking, by not exploring more. There was just a bunch of other careers that I probably was never aware of, where if someone explores a bunch early, say, take a whole bunch of classes in their first year or two of college in different fields and try out different clubs, they're just gonna have a lot more information, so they can more effectively optimize the next couple years or their career after college and reduce the risk that they get out of college, and are like, “Okay, I really like this thing, but for whatever reason, this isn't the career path for me. And so now I have to start over with something else because they never tested if there were other things.”

SPENCER: So for the last topic, I want to talk to you about this idea of how do we have a sustainable work-life? And how do we find happiness and success? What do you talk to your clients about regarding those things?

LYNETTE: This usually comes up when people are asking me something like, “How important is it that I take the weekend off?” And I think that's kind of missing the point. I think it's not about the concrete factors, but about what gives you energy. So when people think of rest, they often come to mind like, “Okay, I'll take a day off and watch TV or something.” And I think that his recovery time – the brain dead low-effort thing. That is what you need when you've just worked yourself ragged, and you don't have the energy to do anything else. But I don't think that taking a week off and going and laying on the beach is going to solve any of the problems of feeling overworked or burnt out. Instead of people focusing on energizing activities, think about the things that when you do, then you come away feeling excited, ready to go and do stuff. This might be a kind of leisure activity. Like, for me, painting is a good thing that I can do when I'm too low energy to write or something, but still leaves me more energized, not done. But also writing or some sessions with clients leave me super energized. And for other people figuring out what things leave them energized, whether this is work or others was, is what gives them that sustainability over time, I make the joke that if you're burnt out from doing a job you hate, it's kind of like an abusive relationship, taking a week off is not going to solve any of the problems in it. But if you can, instead create a work that is exciting, at least some of the time, and gives you that energy that's going to be way more sustainable over time.

SPENCER: So I think what you're saying is that while there is a time for let's say, watching TV like you're you know, you've worked really hard and you're kind of brain dead, and you're like you can't do much else. And it's, you know, kind of relaxing. A lot of what we need to feel good is actually things that don't just, like let us do brain dead activities, but things that actually energize us and motivate us. So they might be work activities, or they could be meditating or could be playing with your dog or whatever. But basically, by the end of that, you're actually feeling ready to go enthusiastic about doing work and that kind of thing.

LYNETTE: Yep. I wrote this book by a coach. He coached people like Sam Altman. It's called The Great CEO Within and he has this idea of an energy audit, where you go through your week, and every activity you highlighted as either something that gave you energy or drained your energy. And looking at this, you focus your energy on the things that are your time on the things that are giving you energy, and his target is that 75% of your time should be spent on things that give you energy including all of your work time. Because that right there is putting you in a zone where you're going to be excited and continue doing this work for a long time. And you're operating in what he calls the zone of genius, which is where you're going to be thinking about it a lot, you'll have the ideas ruminating in the back of your head. I think that depending on where you are in the career, that may be a hard thing to do to aim directly for that. But I think that starting doing an energy audit, and looking at what kinds of activities are energizing, and exciting and which ones are draining energy is a pretty good way to start thinking about sustainability and burnout.

SPENCER: I like that idea. I would add though, I think there is something really valuable about like, let's say taking a week off for vacation. And to me at least I think it's more about having time to look forward to. So you're like, Okay, you know, sometimes work is going to get grueling, like you know, even if overall, you really like your work and find it really energizing. There's just something really nice about it, like, “Oh, I know, in three weeks, I'm going to be on vacation. I'm doing this thing that's really fun.” And then it's like that thing you're looking forward to in the future. I feel like it also gives you energy in a way. Can you relate to that at all?

LYNETTE: To some extent? Yes, I think I feel a lot less of that. I'm much more living in the near future. And so having things that I'm looking forward to later today is going to be more important than a vacation in the future. That's just for me personally, but most of us think that vacations are great. I'm not dissing them, I'm saying that you should not be using a vacation to try and cure being miserable the rest of the time.

SPENCER: Right. That makes a lot of sense. So would you see the value of vacations as being? So for me, I do think I often live in the future, in many ways. So one of the times when I tend to find myself most low is when I feel like there's nothing in the future I'm looking forward to. So for me, things like vacations or just fun plans or you know plans to see friends that I'm really excited about can really help with that. But if the point of vacation is not to like to re-energize us like why do vacations?

LYNETTE: Oh, I don't think so I think it can reenergize I like more of the framing of it as giving your brain a shake-up. This can be useful for new ideas, or kind of as a way to benchmark time. I think that day-to-day life can often blur together into long stretches. And the kinds of things that are very different, and hence send it out in your memory are a good way to punctuate this time. And give us time in a weird way, it makes time seem like it's moving more slowly like life isn't slipping away as quickly. Because you have all these things looking back like okay, here's what happened. And it just feels like there's a lot there in your past. And vacations are one of the things that often have this punctuating effect in life.

SPENCER: Yeah, if you spend everyday kind of working in the same office, doing kind of similar activities, there's a weird sense in which life can all blend together. And you're like, Well, what did I do this year? You know, yeah, like, because so much of his self similar doesn't implant as many memories. I think also, another benefit of vacation besides just the obviously the fun and pleasure, which I think is, you know, it's worthwhile. But beyond that, you know, to this idea of shaking things up, it also could just break your routine and give you a lot more free time. And then you can start to think about your destructure of time and what you can do at a time. It's like if you go from, like, being way overworked where you know, you're working really long days every day. And now suddenly, you have free days. And you're like, “Oh, well, how do I want to spend time in my free days? Maybe I'll try meditating.” You know, and maybe you have trouble making the space for that in daily life. But then maybe if you try it on vacation, you're like, “You know, this is actually pretty cool. I should make time for it in my daily life.”

LYNETTE: Yeah. So I will say that the people who are getting bad at vacations usually aren't the ones talking to me about vacations. Because it's way more often than they're like, “Oh, I travel and now all of my habits have fallen to pieces, and I have to rebuild them.”

SPENCER: Yeah, well, you know, that is a problem. Like, you know, our habits can be really beneficial to us. But it also says it's nice to have to rebuild. Because you can send us building better, I guess it could go either way. But thanks so much for coming on the show. This is really fun.

LYNETTE: Yeah. Thanks for having me. It's great talking with you.

[outro]