Kent C. Dodds (00:01.228)
What's up everybody, it's Kent, SeedOdds again. We're here to talk about product engineering with my friend Alex Hillman. How are you doing, Alex?

Alex (00:07.921)
I'm doing great, Kent. It's really good to see you today.

Kent C. Dodds (00:10.016)
Likewise, good to see you too. So I am really excited to talk with you. As I was planning out the episodes, I just asked general ideas for guests and Joel Hooks came in with a whole bunch of suggestions. He suggested you and that's just so obvious because you and I have worked on launching several of my products and you just have an enormous amount of knowledge in like customer research, product fit, all of that. So.

I think that there's a lot that we can learn from you. So why don't we get a bit of an intro or like fill up our context window on your background and experience and then we can get into how to product engineer.

Alex (00:52.497)
Yeah, absolutely. Well, I mean, as you said, I've gotten to work on a couple of your product launches over the last few years, and it's always been really, really fun for a lot of reasons, not the least of which is how well you know your audience and how much your audience trusts you. And it's clear that that's not by accident. And one of the reasons I've enjoyed working with you so much is because I think you do very well what we

try to teach people to do. And by we, mean myself and my partner, Amy Hoy, we started a company almost 18 years ago called Stacking the Bricks, which is an education company focused on helping mostly developers and designers. Though we have students that are in a wide range of creative skills and even some folks that are more far afield. Like one of our alumni is a race car coach. Like he coaches people how to drive race cars and he's applied a lot of the same stuff that I think we'll talk about to that.

point here being as a different kind of engineering, quite literally, but that a lot of these skills are universal. And the thing that Amy and I kind of locked in on all those years ago is that we were surrounded by people who had some of the best creative and technical skills in the world, people that were working for some of the best startups and big companies, not just in the Bay Area, but all over. And many of them were either challenged or frustrated because when they went to go build or launch their own thing,

Kent C. Dodds (01:48.716)
Yeah.

Kent C. Dodds (02:01.933)
Hmm.

Alex (02:15.857)
things would kind fizzle and fall apart. They would build something that they thought people wanted, but when it came time to sell it, nobody bought. Or in some cases worse, think people would buy, but then it would stop or they would want to make a change and because they didn't know why people were buying in the first place, they were afraid to make any changes. And this is another kind of another failure mode. so Amy and I, while she started a software company and actually started a real world company, I started a

Kent C. Dodds (02:17.998)
Yeah.

Kent C. Dodds (02:37.28)
Hmm.

Alex (02:44.561)
co-working space as a freelance web developer myself, the thing we had in common was not the business that we started, it was the way we started it and then we started with customer research before we were doing any kind of building whatsoever online or off. And in hindsight, you know, I think that for some people that's a more obvious statement these days, they said build what people want. But I think there's a bunch of missing advice in that simple little sentence. And so we started sharing.

Kent C. Dodds (02:46.914)
Hmm.

Alex (03:13.039)
what we had done and why we did it that way so that we had paying customers on day one and not just paying customers because, you know, they liked us and we were popular, but because those paying customers saw what we offered and they said, man, that's exactly what I've been looking for. Yeah, I definitely, I want that. I want to use that as a thing I want in my life. And so we started sharing what we knew and over the course of a few years turned that into a curriculum that we taught. And then over, you know, the last decade and a half, almost two have taught

thousands of students all over the world that core set of skills that start with figuring out who you are best suited to create things for, figuring out what problems they have that they want solved and are willing to pay for, and then doing a research-based approach that doesn't require awkward and sometimes difficult to impossible things like getting time to interview people.

about what their pain points are and stuff like that, whether it's because you're more introverted or you just don't have access to those people. That is a very, very steep hill to climb when you were trying to understand who a customer might be. And so we took a very different approach that was intuitive to us, but unintuitive to others, which is observational research. We can talk more about that. And it built our own little empire, kind of like you did, applying what we learned to people who have the creative skills but don't know what to make.

Kent C. Dodds (04:14.318)
Hmm.

Alex (04:36.871)
or if they do know what to make, they don't know how to get people to understand it so that they actually want it and buy it. And over the last few years, since Amy came back from some serious health issues, we're able to work with some amazing creators like yourself through our mutual friend Joel Hooks to apply some of these skills at an even higher level to folks who are not just getting started, but who have already done what we effectively teach our students to do, which is to systematically find people, earn their trust by helping them in public.

and build stuff they actually want. And that sounds very easy and straightforward. And as you know, it's not. But that background is really, I think, what brought us to today. And part of what I think of bringing in this conversation is, you know, we're getting asked every day, like, how does AI change what we do, the processes we teach? And I think similar to your perspective, there are things that absolutely change. And there's some real key fundamentals that not only do they don't change, I think they're more important than they ever were.

Kent C. Dodds (05:36.11)
Yeah.

Alex (05:36.295)
And I kind of want to, in our conversation, maybe see if we can find some of the crossover between what are those things in the sort of product engineering, not just the code engineering, but the offer engineering. How do you engineer people's understanding? How do you engineer desire? How do you engineer people to take action? And not in a way that's like scammy or, you know, we're not here to make people do a thing they don't want to do. We're here to help them make an informed choice.

Kent C. Dodds (05:49.408)
Hmm.

Alex (06:04.207)
and do a thing that is in our mutual best interest. So I like to think it's a healthier outlook on things like sales and marketing that often make developers and engineers really itchy. My background like yours is also in software. And so was like, how do I take this stuff and make it not only make sense, also make it so you can feel comfortable in your own skin while you're doing it.

Kent C. Dodds (06:07.202)
Wow.

Kent C. Dodds (06:15.787)
Mm-hmm.

Kent C. Dodds (06:27.496)
man, I think that a lot of people can relate to wanting to be able to like feel comfortable in their own skin. And especially when we're talking about marketing or personal brand or like all of that, people get really kind of itchy thinking about those things. But there were so many threads that I want to pull in everything that you just said. that like one thing that really stood out to me was you and Amy,

met up because you both managed to build something that people wanted and have paying customers on day one. And then you also mentioned like figuring out how to have those paying customers on day one, even if you're not well known, like you don't have a ton of followers. so, and then you also mentioned observational research. I think these are a lot of things that I really want to dive into. So where do you think is the best place to start? Cause we can go pretty deep on any one of these.

Alex (07:24.401)
Well, I think the starting point for me is always going to be the answer to who and thinking about the audience and stuff like that. And I think that especially as the territory is shifting fast and furious in the direction of everything, AI can do things for us. It can accelerate things.

Kent C. Dodds (07:31.491)
Mm.

Kent C. Dodds (07:41.688)
Yeah.

Alex (07:46.791)
And I think a lot of the questions are around like, where is the room for the human in all of this? And I think at the end of the day, one of the easiest things to remember is that humans are still making the decisions. even to hand off a decision to AI is a decision, right? So I think I want to, let's start by talking about, know, when we talk about audience selection or even understanding who your best audience is, maybe is a useful place for us to start.

Kent C. Dodds (08:02.102)
Yes.

Kent C. Dodds (08:14.176)
Yeah, yeah. So let's say that for a lot of people now, I guess to back up just a second, some people listening to this want to start a company and want to want to build their own product. Some people are working at a big corporation, just nine to five thing, maybe they really love what they're building. But it's not like they need to, they're not necessarily the ones who are looking to find the people to buy the product.

Can the things that you tell us about today, can they apply to those people as well?

Alex (08:48.391)
Absolutely. And if anything, it's another one of those situations where like it's a necessity if you're going to be an entrepreneur. It's an advantage and an enormous one if you are gainfully employed at a company. It's the kind of thing that makes you more valuable. It's the kind of thing that I think makes makes you stand out. It's the kind of thing that makes you hireable. And again, back to the you know, how do we navigate this landscape of AI? It's the kind of thing that will help you be irreplaceable.

Kent C. Dodds (08:54.19)
Hmm.

Kent C. Dodds (09:17.346)
Yeah.

Alex (09:17.691)
And if we talk about this from the perspective of being an employee, you know, I think a lot of folks are in a situation where they get effectively told what to do, right? They're receiving, whether it's full PRDs or you get it at the task level, whatever it is, like you're kind of chomping through other people's assignments. And the truth is, is like, you may or may not be empowered to push back on those things, to ask questions about those things. But even if you're not in

powered to influence them, I think it's still really valuable to know how to look at it and ask questions for yourself. And if you're building a feature, if you are planning out a scope, I was watching a video of you, I think it was on somebody else's podcast about the back button and how like you can break the back button or how often the back button is broken and resubmits. I think the way you said it is like,

Kent C. Dodds (09:54.605)
Hmm.

Kent C. Dodds (10:09.102)
Yeah.

Alex (10:14.233)
If an engineer just thought about going one step further, like it's not hard to complete that loop, but you have to put yourself in the shoes of the person who's using it, what their day is like, what they perceive pain as. And those extra two clicks may not be the biggest pain in their day, but among all the other pains might, you if you can relieve that one, well, why wouldn't you? And I think, you know, you can look at it through the lens of software craftsmanship.

Kent C. Dodds (10:24.034)
Yeah.

Kent C. Dodds (10:36.118)
Yeah

Alex (10:40.955)
which I think you and a lot of your audience do, but even if you're not, think being able to, I think where people feel maybe unempowered in jobs is, they push back on it because of a personal opinion or a personal element of taste. look, your taste is your taste and it's as valid as anybody else's. But a thing that Amy and I have always said to our students, and I think this applies when you are working for a company or you're self-employed, is you are not your customer.

Kent C. Dodds (10:55.874)
Hmm.

Kent C. Dodds (11:10.926)
Hmm.

Alex (11:10.951)
and the taste that you are optimizing for, and this is like weird and counterintuitive, the taste you're optimizing for may not be your taste, it might be theirs. And so getting a sense of what your users care about, what they prioritize, what friction looks like to them may look different for you. And when you start getting not just like a rhythm and practice, but really internalizing what those pain points are,

Kent C. Dodds (11:19.158)
Hmm, yeah.

Alex (11:39.303)
and then you start making your own decisions, which again, may be how you implement a PRD or a task, or it may be how you respond to the assignment and say, hey, I got this, I have a couple of questions. Instead of it making it about your taste or you don't like it, or this is not a best practice, those things can fly over a product manager or a project manager's head. They're like, I don't care, I'm just getting it from my boss.

Kent C. Dodds (12:03.618)
Hmm.

Alex (12:05.603)
One of the best ways to make a persuasive argument when you do not have a position of power is to know who benefits from the decision that you're trying to influence. And when you can speak the language of the customer, you're effectively speaking the language of the business and therefore your employer, and you're able to be a whole lot more persuasive. anything that we could sit here and talk about in terms of like product marketing and sales and how being informed about who your customer is and deeply understanding them.

Kent C. Dodds (12:14.731)
Mm-hmm.

Kent C. Dodds (12:21.909)
Mm.

Alex (12:35.493)
I think is again just as valuable in some cases if not more valuable inside of a company because at a big company scale, you, your tiny action is an enormous lever. It's a fulcrum. You get to impact so many more people with that one small choice or decision. And if you can learn how to be effective in communicating those decisions internally through the lens of your customer, your project managers are going to look at you differently and they're going to listen differently, your product manager and all the way up the food chain.

Kent C. Dodds (12:45.206)
Yeah.

Kent C. Dodds (13:01.197)
No.

Alex (13:03.739)
And you may even find yourself in a position where they're like, hey, like you've been a good programmer engineer, but it seems like you've really got a good understanding of who our customer is. And you may start getting pulled into other conversations. Then that's kind of leads to, again, I can't promise promotion, obviously, but it's kind of the absolutely leads to recognition. And then how you play that recognition in your favor is really the next layer of that game.

Kent C. Dodds (13:14.732)
Yeah.

Kent C. Dodds (13:18.926)
All

Kent C. Dodds (13:26.402)
Yeah. Wow. That this is way more than I expected your answer to be in that resonates so well. And in fact, I have a blog post I wrote a while ago called how to get whatever you want in business or something like that. And basically, yeah, it was pretty, pretty evocative, I think. But I think the basic premise is just, you have to understand how this pushes the mission of the company forward. And, and then

Alex (13:40.743)
Good post.

Kent C. Dodds (13:54.774)
Like if you can't convince yourself that it does that, then don't bother, like change what you want to be what the missionary company is. And so that like that resonates really well with what you just said. Also Aaron Francis on the podcast a bit ago, at one point he mentioned like, if everybody, all of your users are saying they want it this way and you don't want it that way, you are wrong. you got to listen to those.

Alex (13:59.705)
Yeah.

Kent C. Dodds (14:23.118)
those customers. Now, of course, you're going to have users who want it every way under the sun and they're not all right. But really understanding who you're building for is critical. And as you said, you want to optimize for other people's taste. And maybe that matches yours, but maybe it doesn't. And theirs is more important.

Alex (14:43.271)
That's right.

Kent C. Dodds (14:44.514)
Very cool. So getting back to like the beginning here with how do you identify who you're actually building for?

Alex (14:53.159)
So let's keep this in the lane of working within a company because part of the answer here is if you work inside of a company, hopefully they know, right? Or somebody knows. And if they don't, maybe this is a skill you can pick up and find ways to be useful. And again, I've made it through my career even when I wasn't trying to do something entrepreneurial sometimes by just doing a little bit of extra work to just show that I can do my homework. And if it gets picked up, great. And if not, then I learned something.

Kent C. Dodds (15:01.927)
Yeah.

Alex (15:22.087)
But I think within a company where the customer base is pretty well, hopefully pretty well defined, I think the job of the listener in today's case, I think is to get clear on who that is in two ways. One, I think a lot of people get stuck on things like titles and the different terminology of like maybe a role or something like that. And obviously, like if you've worked at more than one company, you know that...

the same title means different things in different companies and contexts, right? So I don't think that titles aren't bad. They can be a useful shortcut, but I think the more useful thing to think about is how do the people in your audience self-identify? How do they describe themselves? One of my favorite, like, and this is like kind of hokey in terms of being a little bit of a trick, but think of it in terms of if that person were to introduce themselves at a cocktail party, not a professional event.

Kent C. Dodds (15:53.169)
Mm. Mm-hmm.

Alex (16:19.847)
but a cocktail party where you just have like mutual social friends. What might they say if you ask them what they did? And, you know, past the cursory platitudes, how might they describe themselves in the first or second turn of that conversation? I think it tends to be illuminating because it's not just what their title is, but it also implies or gives you lots of clues about like what that role is in relationship to the rest of the people they work with.

Kent C. Dodds (16:34.606)
Mm.

Alex (16:50.063)
what their worldview about that role is. And by worldview, I mean, how does the position they hold or the role they hold shape the way that they see the world around them? What is their perspective of truth, right? And if they're a pretty junior person, they're going to see the world one way. If they're a fairly senior person, they're going to see it another way. If they're a junior person who has transitioned from a previous industry where they were senior, but now they're a beginner, that's going be very different from both of them.

Kent C. Dodds (16:50.21)
Hmm.

Alex (17:19.591)
And I'm just talking about duration as one vector for this, but it could be so many others. It could be the scale of the company or the team they're on. It could be what kind of customers or industries your product serves. Those are all defining vectors as we're spending more time in the LLM space. I love the idea of a matrix of arrows pointing in different directions. And I think it does a really good job of illustrating that we often think of things like

Kent C. Dodds (17:20.27)
Mm.

Kent C. Dodds (17:44.376)
Hmm.

Alex (17:48.433)
who is an audience is a very binary singular definition when in fact I think it's more of a directional definition and where it comes from matters more than in some cases what it is. So how do they describe themselves? And again, multitude of answers there. And then the second thing to ask yourself is, where can you observe them in their natural habitat?

And this is when I get back to what I was saying earlier about a lot of times, you know, people think about customer research, they think about surveys. I'm not anti-survey in any way. I surveys can be super powerful, especially if you're serving people who are already customers. But if you're serving people who aren't customers about their problems that you may want to create paid products for, it's a pretty bad indicator, broadly speaking, depending on what kinds of questions you're asking. And most people are not experts at asking survey questions. So it's the odds are stacked against you.

Kent C. Dodds (18:40.834)
Hmm.

Alex (18:40.835)
Instead, our approach has always been to start by asking the question, well, where do they gather? Where do they seek information and knowledge? And maybe more specifically, who do they already trust? Who are their trusted leaders of their ecosystem? If you're in the React space, you're going to do a little bit of Googling, you're going to come across Kent C. Dodds' name, right? Product engineering, same thing. In a different space, you're going to come across, there's always a handful of people that...

Kent C. Dodds (18:53.272)
Hmm.

Alex (19:09.471)
They don't own the industry by any means, but they become kind of dominant voices. And the secret here is not necessarily to listen to what those people do and replicate it. I think that's what people often do, especially on social media. I think that's a mistake. For me, the work is in the comments. It's not what do they do. It's what do the people who follow them talk about? What questions do they ask? And again, where do they gather? Where do they seek knowledge from trusted sources? I think this is another thing that

Kent C. Dodds (19:21.506)
Yeah.

Kent C. Dodds (19:31.47)
Hmm.

Alex (19:39.007)
to tie it back into the AI landscape right now, people are somewhat frustrated and fearful about the collapse of blogging and sharing knowledge and courses and all the knowledge ecosystems, so to speak,

Kent C. Dodds (19:56.118)
Yeah. Even open source, like you're seeing people stop open sourcing things. Yeah.

Alex (20:01.837)
And on one hand I get it, on the other hand I think it's super short-sighted. especially if it's because the if the concern is, well people are just going to get the LLM to do it anyway, right? And I think we're still a very, very, very far distance away from people trusting these tools. Using them and trusting them are very, very different.

Kent C. Dodds (20:07.31)
Yeah.

Kent C. Dodds (20:19.47)
Mm-hmm.

Kent C. Dodds (20:28.738)
Yeah.

Alex (20:30.919)
People use tools every day, people use software every day that they do not trust. I'd argue most people don't trust their software completely and for really good reasons, right? A lot of bad software out there, a lot of untrustworthy software. And we know that AI is untrustworthy in a different kind of way. And if anything, my firmly held belief, one of my most firmly held beliefs right now is that people are actually seeking out trustworthy sources more, not less.

I think where you access them has gotten more confusing and frustrating and fragmented. Social media has gotten harder with algorithmic feeds and just the attention game and all of that. And I think people are kind fed up with it. And they go, I'd rather find a couple of people and groups that I trust to be not the arbiters of knowledge, but maybe to be a first pass of vetting. And this is another thing where I think when people

Kent C. Dodds (21:23.694)
Hmm.

Alex (21:27.527)
make the mistake of misjudging their own lack of experience as a lack of expertise. And I love to remind people that expert is the most relative term in the world. Expert to whom? Right? Is there someone two days behind you that doesn't know something you do? In their eyes, you can be an expert if just for a minute in that one narrow lane, right? So,

Kent C. Dodds (21:52.248)
Yes.

Alex (21:55.259)
I think the opportunity right now, again, whether you're trying to be entrepreneurial or you're inside of a company, is to apply that lens and go, how is my knowledge useful and to whom? How is my experience useful and to whom? And where can I meet them? Where can I find them? If you're inside a company, you can practice this among your teammates. You've got teammates who know things you don't and you know things that they don't. Above you, below you, all different kinds of, you know, within the structure. You can practice this safely.

learn what it feels like to be helpful in public, within the safety of your own team, within your own group, within your own division, depending on how big your organization is. And with a little bit of confidence, you can go out into the internet world and find those places and patterns as well. It might not be on Twitter anymore, unfortunately, but it might, if it is, it's probably in the replies.

Kent C. Dodds (22:53.038)
Mm-hmm.

Alex (22:53.317)
I would be looking for ways to filter out all of the bot noise and find the real questions, not the platitude responses, and start looking for patterns in them and go, do I have an answer? Yes, the expert has the answer. And I think a lot of people cut themselves off because they assume there's another expert in the room that's going to answer for them. And I just think that's self-inflicted wound. It's not about you being the smartest or the first. It's just about being generous and sharing.

Kent C. Dodds (23:12.0)
Yeah, that, yeah.

Kent C. Dodds (23:17.315)
Yes.

Alex (23:22.631)
from my perspective. like, there's a lot of good reasons to keep your thoughts to yourself. Let's be clear. But I think one of the bad ones is because somebody else has already said it or will already say it. And the framing that has worked best for me, because I have those thoughts too. Like I've been doing this for a long time and I've helped a lot people like that. Thought creeps in. That doesn't go away. You just get better at recognizing it and managing it for what it is. My personal reframe is how

Kent C. Dodds (23:23.822)
Yep, 100 % agree.

Kent C. Dodds (23:30.19)
That's fair.

Kent C. Dodds (23:36.578)
Yes.

Alex (23:52.91)
much like how selfish is it of me to not help a person when I have an opportunity to in in in 30 seconds to two minutes five minutes maybe 15 minutes of my day how do I help a person with no expectations in return not what does it do for me what does it do for them that can I put that into the world if even if they don't respond the upside for you is enormous because you get the practice and

Kent C. Dodds (24:19.668)
Yes, yes, this is so good.

Alex (24:22.437)
you get to apply that practice anywhere and everywhere. And again, I'm on my short list of very firmly-hugged beliefs. That is a practice that if you apply it every day to everything, not just your work, not just your engineering, I think the way you experience the world will be better and the world will be better for it too. like, is it not just me being like a glorious optimist? I think you can create optimism and possibility around you. It's kind of another version of that whole like luck surface area idea, which I also very much subscribe to.

Kent C. Dodds (24:47.223)
Mm-hmm.

Alex (24:51.247)
is like if you help people in public, you become known as a helpful person. And so long as you also place appropriate boundaries around that, that is an incredible power that nobody can take away from you. And you also don't need somebody else to grant you. You just get it by virtue of doing.

Kent C. Dodds (25:09.902)
That resonates so hard, Alex. That is exactly how I got to where I am, is by just being that helpful person. There's so many really great points from that, Alex. So one other thing that occurred to me is that I'm always nice to LLMs. I say please and I don't always say thank you because I know that like that's just the waste of a turn. sometimes.

Alex (25:32.529)
But do you do the thing, I do this in the zone, like, when it actually kind of goes either above and beyond or when I'm pleasantly surprised, I'm like, yo, that's awesome. Like, will definitely, like, not, it sounds quite the same as thanking it, but I will definitely give it like an affirmation in that way, which sometimes I like catch myself in like, this feels like a weird thing to do. And I don't think it like affects performance in any way, but I do think that not doing it erodes the habit of how I interact with people. And I think that's why it's important.

Kent C. Dodds (25:40.139)
Ahaha

Kent C. Dodds (25:45.291)
Yeah.

Hahaha

Kent C. Dodds (25:57.43)
Yes, yeah, exactly. Yes, 100%. And so it's what it does to you. And so for the exact reason that you said, you help people not be, you help them to help them, but even if they don't respond, it's fine because what it did to you, it gave you practice, it turned you into, like you see yourself as a helpful person and that's good for you. And it like,

the more you do it, the more people see you as a helpful person. And I really liked that you said you can do this within your own company as well. You can almost think of it as I'm going to develop a quote unquote product of being a helpful person or like practicing with my coworkers. They know that I know this and that I'm willing to be helpful to them and it just opens up a lot of opportunities to you. At the start, yeah.

Alex (26:50.983)
I'll throw one more thing just on the stack there is like, because I know that like a big part of your world and a lot of listeners world is open source. And my background before I got into all the things that do now was open source as well. Like my introduction to web development was deep in the world of open source. And I think the values of open source when they are held thoughtfully and truly are pretty special and are worth like owning and

I don't know, I don't want to say defending, but certainly considering, right? And I think what we're really talking about is like open source, but for other things. Like how do you apply an open source mindset and mental model to things that aren't code, but other kinds of knowledge? And my understanding of sort of, even without getting into the licensing legality of open source, the spirit of open source, as I always understood it, is I'm going to put this out into the world with no expectations.

Kent C. Dodds (27:27.66)
Hmm.

Kent C. Dodds (27:48.524)
Yeah.

Alex (27:48.955)
My only hope is that if you improve it in some meaningful way, that you contribute those meaningful improvements back and then we have an option to fold them back in so everyone else can benefit. If you apply that mental lens to everything else you know and do, that's really all we're talking about here. It is what do you know, what do you think, what are your experiences? And they don't even necessarily even need to be entirely yours. Like you can also be...

Kent C. Dodds (28:00.525)
Yeah.

Alex (28:16.315)
before I was saying, who are those trustworthy people in the front of the room? Sometimes those people become trustworthy, not because of their own opinions, but because of their judgment and taste of other people's opinions. You become a valued filter or a magnet. would say, good people attract good people. Good people also filter good information. And I think it's all different versions of the same thing is kind of why I bring it up. Because I think open source as a philosophy, I think, clicks with engineers.

Kent C. Dodds (28:25.326)
Hmm.

Yeah.

Alex (28:46.075)
because you've experienced in many cases the full spectrum of it, like the best of open source and the worst of open source. And so I think you can apply all of those heuristics to the things that we're talking about here and realize that you're not going in from ground zero. You actually already have some muscle for this in a thing that is, you know, previously in code. Now it's like, well, what else could I apply this to?

Kent C. Dodds (28:52.675)
Yeah.

Kent C. Dodds (29:09.302)
Yeah, resonates as well. Earlier you were saying like, when you decide who you're going to be building products for, you just imagine in your mind, how would they describe themselves? And I imagine that the way you identify the people you go through that figurative experiment with is just like, you identify some sort of problem that exists in the world, whether you have it yourself or not, like you're

Second cousin has this problem. You're like, you know, I'm gonna build a product for that and that's how you identify Who is your audience or is there another more scientific way of identifying who you're trying to solve a problem for?

Alex (29:50.471)
That's a really good question. I'm really glad you asked it because I think that the example you gave is one of the more common ones that feels like it should be right. And in part because it can be, but I feel like we're dealing with a lot of survivorship bias because you can follow the exact same path. And I'd argue 99 times out of 100, it's going to be a failure. So what's the difference between the 99 failures and the one success? And how do you create more of the success categories, the question that we ask?

Kent C. Dodds (30:17.793)
Yeah.

Alex (30:18.951)
The approach that we've taught and kind of refined over the years, and I should say that like the base of everything that we do, we teach it prescriptively. Like I believe you do too, so that you can learn it, understand it, internalize it, and then adapt it and get fancy, right? So I want to like pre-seed with that, because I'm going to say something that sounds very prescriptive, and engineers are often very resistant to it, but I want you to use it as a mental model first, and then we can talk about expanding it. So there's basically,

Kent C. Dodds (30:36.246)
Yeah.

Kent C. Dodds (30:43.383)
Yeah.

Alex (30:49.319)
three versions of an audience that every single professional has built in. So I'll start by number one. have our focus, our students focus on serving other professionals. Now know a big part of your background and a lot of your audience focuses on consumer product. And consumer product works when you are at scale and typically in order to get to scale you need one of two things. A giant pile of money, which is like kind of

Like you can acquire a giant pile of money if you know what you're doing, or you're very lucky, which is a lot harder to acquire, right? And often luck is, or I should say wealth is misattributed to luck or the other way around. So whether it is consumer or business oriented, we would encourage individuals to focus on professional audiences if you're trying to build a thing to sell, just because it's easier to sell, especially if you're new at it. But in terms of the product conversation, it doesn't matter.

Kent C. Dodds (31:19.638)
Yeah.

Kent C. Dodds (31:25.901)
Yeah.

Alex (31:47.651)
as much. So whether we're talking about a professional audience or a consumer audience, the next question is, is it an audience that you belong to? And folks are often really resistant to this for a variety of reasons. Devs in particular are like, well, devs are terrible customers. I know because I am one. And like, there's some truth in that. But the thing that I also like to remind folks of is

Kent C. Dodds (31:56.558)
Mm-hmm.

Alex (32:12.473)
Whatever relationship you have with other developers is based on the kind of work you've done and the kind of interactions you've had with them. Whether those developers are your peers, your subordinates, or senior to you, that is a defining relationship. When your relationship with that group is based on you being a trustworthy source of help, the relationship dynamic looks very different. And so I always encourage folks, if there is a resistance to a group.

Kent C. Dodds (32:35.725)
Hmm.

Alex (32:40.091)
Like really introspect why that is. Is it because you had so many terrible experiences with them that you absolutely cannot imagine spending another minute with them? Then you'll probably pick a different group. But if it's just like I'd rather not or I'm more interested and passionate about this other group, I'd say if you're learning, if you are new to this, doing this with a group that you do not belong to is not impossible. It's just like a thousand times harder.

Kent C. Dodds (32:51.955)
Haha

Alex (33:08.549)
And it's kind of like trying to learn, you know, it's trying to learn, like trying to learn how to pass your driver's license test in a, an F1 race car. Like, why are you playing this on the hardest mode? And then also, how do you know whether or not it's because it didn't work or because you're just a newbie and not good at it yet? You just don't have the reps in the practice yet.

Kent C. Dodds (33:18.914)
Hmm. Hmm.

Kent C. Dodds (33:32.216)
Can I interject? know you have the third thing, but what if you get a partner who is in that world? So like, I'm not in the world of scheduling dentist appointments, but this guy has, he's been doing that for a long time. He's a dentist or he was a dentist and he knows that world really well. And we were gonna build a company together.

Alex (33:35.302)
Yeah.

Alex (33:54.951)
So I would say it can work. The risk that you bring in is you, what you can't do in that scenario is you can't proxy through his brain. If the goal here is for you to get the understanding, you can, you could, you and that other person could go through the training that Amy and I offer to our students. And even if you went through the identical training, your takeaways would be different because he's got lived experiences that you do not.

Kent C. Dodds (34:06.316)
Hmm, well tell me more.

Kent C. Dodds (34:22.2)
Mm-hmm.

Alex (34:24.903)
Right. And so if your goal, which the context of the conversation is to deeply understand a customer, every layer of abstraction that you add between you and the customer, starting with another person, other versions, this could be other people's research. Are they bad? Of course not. They're just not. I think that it's sort of like the difference between like 1080p and 4K. Right. The picture is still there, but don't you want the highest resolution you can get?

Kent C. Dodds (34:48.77)
Mm-hmm.

Alex (34:53.071)
I want the full understanding. want nature channel in 4K. I want to see every hair move in the wind because that's where the understanding comes from. It comes from the nuance and the detail. And when you have to proxy it through somebody else's pre-existing understanding, better than doing it yourself, a thousand percent. But it's still not going to get you the understanding quite the same way. you know, is it, is it,

Kent C. Dodds (35:05.07)
Mm.

Kent C. Dodds (35:18.858)
OK, so this is

Alex (35:20.815)
Is it doable? Absolutely. And if it's the only path in that you have by all means, and if it's what works within the company that you work within, I feel like that's in an environment where you're the customer of the company you work for probably is different from who you are. I think that's a case where let's not perfect be the enemy of good here and say, get that person on your team, but now your challenge is not extracting knowledge from them. It's sort of,

Kent C. Dodds (35:37.133)
Yeah.

Kent C. Dodds (35:43.01)
Mm-hmm.

Alex (35:50.375)
I feel like I'm throwing a whole lot of metaphors out here. But I think it's sort of like the equivalent of when you do like tandem skydiving. Like I think what a lot of people would do is I would go up in the plane and we'd jump out side by side when what you really need is like be grafted onto that person. And really that other person is doing the dive and you just kind of get to watch through their eyes. I think the best, if you do not have the domain,

Kent C. Dodds (35:53.845)
Ha!

Kent C. Dodds (36:02.168)
Hmm.

Kent C. Dodds (36:09.272)
Yeah.

Alex (36:17.839)
It's not just about the knowledge also. Sometimes it's about the trust. Back to take this all the way back to trust is like, if you're proxying through somebody else, nevermind the speed of it.

Kent C. Dodds (36:21.196)
Hmm.

Alex (36:31.271)
How is the person who you are observing?

If they don't know who you are and you are in the mix, people's behavior is going to change, right? They're going to be a little perhaps more guarded. Maybe their answer is to be little less complete. And frankly, therefore both of you, the person with the expertise and you, will be missing nuance. But the problem here is you don't know what you're missing. So again, I want to let perfect be the enemy of good here. And again, a lot of folks that are listening are in an environment where you and the audience

Kent C. Dodds (36:47.05)
Mm.

Alex (37:06.449)
for whom you build are not the same. I think the question I would be asking is, who are the people who do know? And how can you build a relationship with them that helps you get in the room with the people, right? So treat that person as a channel to get to those people, but the fewer layers of abstraction you can move through, I think the better your understanding ends up being.

Kent C. Dodds (37:20.622)
Mm-hmm.

Alex (37:34.311)
And for the sake of this conversation, feel like that is the thing we're optimizing for is your understanding, how well you understand those other people. The bar is really low. So every version is going to be better. And also again, while you're practicing, practice through those people, practice with those people. Find the person who I think about some people early in my career who I kind of persuaded to let me go on a ride along.

Kent C. Dodds (37:46.306)
Yeah, yeah, that's true.

Alex (37:59.527)
effectively where I was like, I'm going to to this meeting with you. I'm going to be, I'll be silent. I'm not asking any questions. I'm not answering any questions. I just want to listen and learn. And I feel like if you can find ways to be in those rooms and go in and be clear as like, I'm here to listen and learn. And then maybe afterwards ask them, Hey, the, you know, the, the, were doing, you know, as a client or a customer or whatever it is. Um, I heard them say this, you know, can you explain to me what that means?

I think is a good way for you to get some training wheels on this process before you have the direct connection yourself.

Kent C. Dodds (38:28.578)
Mm-mm.

Kent C. Dodds (38:35.392)
Yeah, yeah, that makes sense. you're kind of saying like, for if you're just getting started in into developing these skills, then being in the audience that you're building for is kind of like a write what you know sort of thing. Like you, you don't have to worry about the proxy conversations that need to happen or whatever else, because there are so many skills to learn that you don't want the extra weight of having to like figure out

what matters to that group of people.

Alex (39:07.307)
I think that's right. Like I said, at the beginning, think the thing to know and to remind yourself of is it's very hard to know if what you're doing is working or not. The feedback loops are slow and sometimes ambiguous. And I think that's the reason why say those layers of abstraction make it so that to take it back to the high definition TV metaphor, it's like maybe you've got HD but the audio is a little bit out of sync.

Kent C. Dodds (39:18.747)
yeah.

Kent C. Dodds (39:24.866)
Yeah.

Kent C. Dodds (39:37.165)
huh.

Alex (39:37.231)
It's just to kind of mess with your ability to interpret the entire message, meaning all of those kinds of things. And so again, like maybe the answer here is not starting with your customers, but instead starting with groups within your organization where you can get practice with this muscle where not only the risk is lower, but there's just like fewer layers for you to try and parse through and get a sense of, when I see this,

Kent C. Dodds (39:53.336)
Hmm.

Kent C. Dodds (40:01.997)
Hmm.

Alex (40:06.223)
my understanding and I interpret it this way, my understanding is reliable and I can then deploy my understanding in the useful ways that we've talked about.

Kent C. Dodds (40:14.54)
Yeah, okay. So you were in the middle of explaining like three steps or three ways to mental models for identifying who you're building for. What was the third one?

Alex (40:25.541)
The third piece is sort of a natural kind of bifurcation, right? So people go, well, I was supposed to pick people who are like me, that is, know, professionally or hobby or whatever it is. Then there's sort of three layers within that. There's the true beginner, is the person who's just getting started. They don't know what they don't know. There is your peer. There is the person who does what you do and wants to do it more like you, right? So.

Kent C. Dodds (40:53.486)
Hmm.

Alex (40:54.299)
there isn't necessarily an experience level, maybe because of the approach that you've taken or whatever it is or the insights that you've gained, you get to help the people who are positionally kind of lateral to you. And then the last one, and this one is perhaps the most interesting for this audience, is people who hire you. So again, this is a thing where you can not only apply this to your product engineering, you can also apply it to applying for jobs and things like that.

You can apply this process to who are the people that are making decisions about who gets brought onto a product team. What are their priorities? What do they care about? What do they perceive as painful? What are they optimizing for? What kinds of people do they and do they not want on their team? What kinds of experience do they want their team to be known for culture wise? What do they want their product to be known for? Because I watch people talk about products and one of my favorite things, the fact that we live in a world where massive products get used by

not just thousands, but millions of people, and you can know the name of a person who made it, I think is really cool. Right? And I think that's an intentional choice within some organizations and not others. And so when you, again, take a look at who it is that you are engineering for, again, whether it's an end user, a teammate, a boss, an employer, what are you engineering for? I think the question becomes, well, what are they engineering for?

Kent C. Dodds (41:59.522)
Yeah, yeah.

Alex (42:22.951)
And how do you align your interests with what they're engineering for? I think that's the lesson that applies through every version of this, whether we're talking about software engineering, experience engineering, lifestyle or job engineering. think these are, at the end of the day, it's all intentional design choices in the direction of a desired outcome, where there is another party that needs to be involved, and we need to be on the same page so that we're moving in the same direction and we...

are achieving shared goals together.

Kent C. Dodds (42:54.56)
Yeah. Yeah. I like what you said, you know, focus on what they are engineering for. it, that makes sense in the context of if you're, what you're building is for engineers. But if we broaden that to what you're building is for as a consumer, even a non-technical professional, that makes me think of the, concept of the job to be done and just like really understanding why, like what are they trying to actually accomplish and how can your product

Alex (43:14.726)
Yes.

Kent C. Dodds (43:24.482)
kind of slip in there to optimize that job to be done for them.

Alex (43:28.945)
sometimes it's so confusing and this is where like working in fields that are not your own gets so weird so quickly is you come at it from your worldview. Like one of our go-to examples is the number of times that people have said I'm gonna build booking software for salons. The salon or the barbershop I go to like they use this old paper book and they're sliding paper clips and all these things and like honestly it was pretty wild but like the pay took a pandemic. It literally took a pandemic for

for that to shift. For decades, people have been trying to get salons to adopt this stuff. And the first time it took movement was literally a global economic shutdown. It shows you how resistant non-technical people are to technology adoption. Because from their perspective, the way they see the problem is not going to be the same way you see the problem. You may even be looking at the same problem, but what you believe is a potential solution.

Kent C. Dodds (44:01.858)
Yeah.

Kent C. Dodds (44:09.633)
yeah.

Kent C. Dodds (44:22.542)
Yeah.

Alex (44:28.357)
is going to be wildly different. So what you're engineering for is not just the outcome. That would be easy. Everybody could do that. What you're engineering for is what they believe the path to the outcome is. And to a degree, there's multiple layers to engineer there. if you're asking them to do a thing that is different from these two, you also have to engineer the trust. I've got to give you the thing you want, just enough to get you to be willing to follow me in this slightly different direction that you weren't going.

Kent C. Dodds (44:35.33)
Yeah.

Kent C. Dodds (44:39.822)
Hmm.

Kent C. Dodds (44:55.519)
Interesting.

Alex (44:57.423)
And if you can't do that, then your job is really to meet them where they are. And in order to do that, you need to understand how they believe the steps and the workflow works. And it's in the most confounding part, Kent, I think, is it is often not logical. It is not sequential. is irrational and sometimes like self-defeating.

Kent C. Dodds (45:01.314)
Mm.

Kent C. Dodds (45:18.435)
You

Alex (45:22.439)
And the hardest part to remember is in that moment, it is not your job to change their behavior. It is to help them get the thing they want done well enough that you have an opportunity to change their behavior, is the way I think about it.

Kent C. Dodds (45:22.508)
Hmm.

Kent C. Dodds (45:27.534)
Hmm.

Kent C. Dodds (45:36.058)
yeah. Yeah, that is so good. I have people in my life who are not technical and they say they want technology problems and they're like, I know it's better to do it this way, but they just, they don't. it doesn't compute for me, of course. So that resonates really well. And so it almost feels like in that setting, you have to meet them where they are to almost trick them into.

changing their workflow to something that you do know is better for them, but is approachable for them to even get started on the path. And so maybe what you build, for them is not like the end results of like the perfect utopia that you see, but you build a stepping stone that once they get there, then you can say, and now we've got this and we just slowly evolve their process into utopia.

Alex (46:18.567)
Exactly.

Alex (46:29.223)
And it's kind like when you get into the pool and it's like, oh, this is a little cold. And you kind of inch your way in and then you really just need to get those first couple of dips and then you acclimate. think if you think about your product, and I think good product onboarding does this. I think we think of product onboarding as taking people through all the features that are available to them. The conversation you had with Dax about progressive disclosure of features I think is another version of what we're talking about here.

Kent C. Dodds (46:47.394)
Yeah.

Alex (46:56.059)
what are the things that they know they need and are looking for that earn me the ability to slowly introduce them to things that they may be looking for, maybe not, but the real, I think the magical experiences that you and I and the listeners get to make are the ones where people feel anticipated. Not like bullied into using a feature, which I think is a thing we see all the time, but instead they go, while I was using this thing that I find useful.

Kent C. Dodds (47:17.496)
Hmm.

Alex (47:26.043)
This other thing kind of popped up over here and it showed up at just the right time where it's not interrupting me, but it is related to what I'm doing. And maybe it just plants the seed that I'll check it out later. Or maybe the thing it offers is the right thing at the right time where I'm like, yeah, I'm used to doing it this way, but I'm willing to take this slight detour, this little side quest and explore something new if you can communicate how much time that's going to take or how much effort or what's expected of me.

Kent C. Dodds (47:46.958)
Mm-hmm.

Alex (47:55.947)
I think that everything we're talking about here doesn't just start at the beginning of a product or a feature. It's the entire experience of what the user has either signed up for. Again, if we're talking about large scale software, often they didn't even sign up for it. Somebody else chose the software for them.

Kent C. Dodds (48:03.042)
Mm-hmm.

Kent C. Dodds (48:12.704)
Yeah, yeah, you probably hate it that they have to do this and yeah.

Alex (48:17.371)
I think a lot about that is like, can I create a magical experience for somebody who is in the moment, the most resistant to it? That's not necessarily a thing you should optimize for, but I think it's a useful mental exercise to think about it and go, okay, if that's the hard version, where's the law hanging fruit and how do I get some practice there?

Kent C. Dodds (48:23.137)
Mm-hmm.

Kent C. Dodds (48:34.367)
Yeah, and it's more than just like thinking, you know what would make this better is sound effects or something, but like actually talking to those users who, or being in the room, like you said, to understand why they're resistant and what you can do to make that experience at least less painful, if not delightful.

Alex (48:54.257)
Yeah, a little poof of confetti or an animation, well, I find delightful when done well and tastefully, is not the same as feeling anticipated. And I feel like one of the reasons why people like the experience they have with the chat style interface with LLMs is it has, I don't want to call it the ability because sometimes it's more of a simulation than a reality and sometimes it's hard to the difference.

Kent C. Dodds (49:04.205)
Yeah.

Alex (49:23.707)
sometimes feels like it's able to anticipate what you wanted to do next, right? And like I built into my personal AI executive assistant, I know you're working on Cody and I want to trade notes on this at some point. But a thing that I built into mine is I have a little bit of code that gets called at the end of every agent term and it looks at the recent transcript and goes, is there either a yes, no question or multiple choice question? And if it is, it just generates

Kent C. Dodds (49:27.105)
Yeah.

Kent C. Dodds (49:33.666)
Mm-hmm. Yeah.

Alex (49:52.519)
So instead of me having to type those things, I just tap the button. And I have that both in my custom built UI, but I also have it in my Discord UI. And it's one of the things that I could very easily take for granted right now because I use it so much every day. But it was an opportunity to build a thing for... This is a weird thing. To build a thing for myself that still delights me because unlike most software where I built it and so I know what it's going to do every time.

Kent C. Dodds (49:53.87)
Hmm.

Kent C. Dodds (50:07.746)
Mm-hmm.

Kent C. Dodds (50:14.892)
You

Alex (50:20.699)
Like it will often generate buttons that I didn't tell it to generate because it's generating buttons based on the context that came before it. And the buttons will be like, yes, that's actually exactly what I was going to tell you to do next. And feeling anticipated is like, it's like a hug from software, I think. And that's what makes good like marketing and sales too is feeling tricked sucks. Feeling understood and anticipated.

Kent C. Dodds (50:26.828)
Yeah.

Kent C. Dodds (50:38.11)
my god.

Kent C. Dodds (50:45.325)
Yeah, yeah.

Alex (50:48.857)
is like one of most human desires in the world, I think. And I think the power that software engineers have, perhaps uniquely in the world, is the ability to create that feeling for so many people at scale. I think that's a really cool superpower that I, like one of my hopes for the future of software is that more people are able to do things that allow their users.

to feel that hug of, you thought of me. And what's cool about that example is sort of this interesting tandem, you thought of me, where the you thought of me is both the engineer that built the feature that allows the LLM to do it, as well as the LLM using its predictive capabilities to do its own kind of anticipation. That's a marriage between human and software that we haven't had before. And I think opens up a ton of very interesting possibility that I haven't seen explored as much as I think it can and will be.

Kent C. Dodds (51:47.148)
Yeah, yeah. And I agree that feeling tricked is we don't like that, but even more subtle things, feeling patronized or patronized too as well. So confetti after deleting an issue or something, like, okay, maybe, but like often, yeah, exactly, exactly. And at that point, it's just like, is this like a little pat on the head? It just doesn't feel good. you do have to...

Alex (51:58.391)
yeah.

Alex (52:05.159)
It was cute the first time.

Kent C. Dodds (52:16.75)
Um, find where that line is between, uh, being act actively helpful, um, delightful to use and, um, and anticipating the needs, um, and also being patronizing or, or tricking. So, um, Alex, man, we, we barely got through, like, how do you identify who your audience is and how do you, you know, uh, like get into their mind space and mental model?

Didn't really talk too much about observational research and specific techniques. Do you have a second to talk with us a little bit about that and maybe some next steps once you've identified who you're working for?

Alex (52:55.879)
let's do it. We can do a little bit of a speed run here. So, I mean, I think the starting point here is to sort of do the contrast between the survey and like the survey approach and the observational research. I mentioned like the Nature Channel in 4K. The research methodology that Amy and I developed, really Amy invented and we've refined over the years, we call sales safari. And the reason we call it sales safari is as a contrast to what I think most people do, which is

Kent C. Dodds (53:07.395)
then

Alex (53:26.203)
going to the sales zoo. It kind of falls apart pretty quickly, but stay with me. And I think when most people treat research as how do I find this kind of confined and contrived environment of sitting down with a user or a business owner or whatever it is or a teammate and asking questions. And it's not that you can't get useful insight. Again, I want to be clear that I want to paint this as a binary. It's a spectrum.

Kent C. Dodds (53:30.582)
Okay.

Alex (53:53.137)
But I think there are some key limitations to that. One is your skill at asking good questions. Most people, frankly, aren't good at it. You don't know what to ask, and therefore you're not going to know how to ask. The other part of it is, I think this comes back to trust in some cases, but also just like the nature of humans is people's answers are not always honest or not complete. Exactly, they're not lying to you.

Kent C. Dodds (54:04.845)
Mm-hmm.

Kent C. Dodds (54:15.84)
Yeah, because they don't know.

Alex (54:19.877)
they don't know or they don't want to hurt you or they're hedging or the way you asked maybe pulled them in one direction but what you really wanted was another. So the way people answer, yes, 100%. Yeah, Rob's book, The Mom Test, we're huge fans of and mutual fans with that author and that book. Rob's one of the best at the job. So observational research.

Kent C. Dodds (54:29.305)
It's a little bit like the on test as well, right?

Alex (54:49.179)
we tell people to go on the internet and find their audience. Now, again, this is a thing that has changed and in some ways has gotten harder. If we go back not even 10 years, when we started it was go on internet forums and then it was email lists and then it was social media and the comments and now it's Discord and Slack and everywhere. The conversations are fragmented all over the place. And the truth is, it's harder but that doesn't mean

Kent C. Dodds (54:54.552)
Hmm.

Kent C. Dodds (55:15.106)
Hmm.

Alex (55:18.663)
to not do it, right? When we started doing this, we'd say, go to those places, find people that are talking, look at what questions they're asking, look at what problems they're talking about, and people would say, they'd come back and they'd say, I looked at it, I just saw a bunch of people complaining. And I'd say, that's what you're looking for. And what we realized is people didn't know what they were looking for and they didn't know how to note it. And so the encouragement that I would give you is to go to these places and you're looking for four key things.

Kent C. Dodds (55:19.993)
Kent C. Dodds (55:32.187)
huh. Yeah, that's good sign.

Alex (55:48.069)
as a straight from the Stocking the Bricks 30 by 500 playbook, we go in depth in our material on how to do this, but I'm give you the speedrun version of it. Number one, and probably the easiest thing for you to surface before you even get to pain is jargon. You're looking for terms, terms of art, language. This goes back to a version of how people describe themselves. What are the terms that people use to describe their work, describe their problem, and start taking notes on it?

both easier and in some cases harder because if you are in the audience, you also may not recognize those as terms of art. You'll just recognize them as words, right? And so you kind of have to of step out of yourself for a minute and go, I'm not going to be an engineer for a second. I'm observing engineers. What words are words that only engineers use or this kind of engineer or this kind of end user, whatever background, customer, what goal they're trying to do. You're looking for the terminology. You one of my favorite examples is, I think I even said it in this conversation is like,

Kent C. Dodds (56:35.906)
Hmm.

Alex (56:47.569)
customer versus client. They both describe very close to the same thing. But if you talk to a freelancer, you describe their client as a customer, they're going to look at you kind of sideways and maybe even question whether or not they should trust you. Right. So the old

Kent C. Dodds (57:03.434)
Yeah, it's like you're doing the three versus three thing.

Alex (57:07.175)
That's such a good example. Yes, exactly. So jargon, easiest place to start. Also like a lifetime of practice to like start picking out jargon. The second thing you're going take notes on is pain and problems. There's really two layers of that. One is, and sometimes the easiest one is the question, but what I would challenge people to do is when you see a question on the internet, to ask yourself, what was going on in this person's head and in their day?

in their life that made them feel a problem to the degree that they went on the internet and asked a internet full of strangers for help. The question is not the question. The question is the answer to that that I just asked. And that is a challenging mental exercise, especially at the beginning, but it is one that gets much easier, much faster with practice. And the beautiful thing is, is like, you don't need to be right. The goal here is not to be correct. This goes back to the vector thing. It's about where does that direction pull you?

Kent C. Dodds (57:44.717)
Huh.

Kent C. Dodds (58:01.602)
Hmm.

Alex (58:02.087)
Right? And so you're looking for those, the subtlety of those pain points. Another thing you can look for is what's not being said here. Cause that's a really, like the absence of information is its own kind of pain. What, not just what are they asking, what are they not asking? You know, how often are we saying, why aren't you reading the docs? For instance. And a really good answer is, besides people being lazy, of course, is the docs aren't good. Right?

Kent C. Dodds (58:26.734)
Ha

Alex (58:29.671)
or the docs are out of date, or my experience with the docs is that they're not good and they're out of date, those kinds of things. And that's a subtle difference, but an important one. So we've got jargon, we've got pain. We've got the third one we talked about, I would argue is one of the hardest, is worldview. And that is, what does this conversation suggest about the way they see the world? And one of my favorite ways to figure out, you don't have to necessarily pull out the worldview of an individual, but if you look at, here's my favorite thing, go on Twitter right now.

Kent C. Dodds (58:34.382)
Hmm.

Alex (58:59.591)
It's not hard. Find an argument. If there are two opposing views that could theoretically, you don't have to agree with both. But if both could be functionally true at the same time, what you're really dealing with is not a disagreement. You're dealing with a misalignment of worldview or two different opposing worldviews. Apple versus Android is like the quintessential example. Functionally identical, but they are catered to.

Kent C. Dodds (59:02.37)
Yeah.

Kent C. Dodds (59:20.162)
Mm-hmm.

Alex (59:29.199)
a worldview of I want to tweak, I want to optimize, I want to control, want visibility transparency. With Apple, least hypothetically in the past, I just want it to work so on and so forth. So worldview is the third piece. And the fourth one, and this is one that actually is like the most on the surface and gets also missed the most is what do other people recommend? So again, one of my favorite things to go on the internet, when I'm trying to understand people, I look for questions.

Kent C. Dodds (59:40.429)
Yeah.

Alex (59:57.595)
And then I go straight to the comments or the replies or whatever it is and I go, what are other people recommending and why? Because you can get information about what kinds of things they love, what kinds of things they hate, what kinds of things they share, what kinds of things they distrust, what kinds of things that they see it and they run the opposite direction. So recommendations are full of clues for you about what people actually want and why.

Kent C. Dodds (01:00:03.576)
Hmm.

Kent C. Dodds (01:00:19.608)
Hmm.

Alex (01:00:27.687)
And so those four categories give you a ton of information. You can go to any forum, you go into a Discord, pull up a thread of conversation, a bunch of back and forth, go on Twitter, anywhere where there's any kind of back and forth commenting happening, and start practicing looking for those things. And the thing that sounds silly, unless you are a recent college student perhaps, is the importance of taking notes here. And this is not...

This is not from like a teachery type assignment. This is about how the information gets into your brain and stays there. And a lot of folks would go through and I mean, right now people are like, this sounds so easy. I'll just have an LLM go do it. I mean, you can, but you're going to, similar to, think that's the worst version of the proxy problem that I was talking before going through an employer is you deprive yourself of the process of using that information in your brain where it goes from information to understanding.

Kent C. Dodds (01:01:02.509)
Hmm.

Kent C. Dodds (01:01:09.718)
Ha

Kent C. Dodds (01:01:17.058)
Hmm

Kent C. Dodds (01:01:26.572)
Hmm, yeah.

Alex (01:01:26.591)
The LLM can simulate understanding, but it's got to get in your brain. Your brain's going to put it together, and that's where the understanding comes from. So taking notes is an important part of that. And like the most extreme version of this, I'd say sometimes people go and they'll go in, they'll just copy and paste it into a Google Doc or your VS code, your text editor of choice, whatever it is. And I would strongly encourage to actually type it out. Or if you're doing voice these days, talk it out.

Kent C. Dodds (01:01:53.08)
Mm-hmm.

Alex (01:01:56.195)
is better than copy and pasting it. And there's something, I'm sure there is a scientist out there that can explain this. The difference between how knowledge is actually not just gets into your brain, but retained. I know this is like the way you teach through practice and exercise and repetition. Watching the information is very different from doing the thing and doing it repeatedly. Typing it out or talking it out gets it into a different part of your brain for storage.

Kent C. Dodds (01:02:18.455)
Yeah.

Alex (01:02:24.337)
for recall and for synthesis. And that's just the part that happens in your brain. I haven't talked about the last part, and we can wrap up on this for answering this question, which is think of this entire thing as a loop. So you're going out, you're collecting all of these things. It's valuable to take those notes because you're going to be exhausted at the end of that session, and when you come back tomorrow, you're going to have forgotten a large portion of it. So the notes are there to support future you, and when you come back to it tomorrow, you're going to do two things. One is you're going to review all the notes that you took.

Kent C. Dodds (01:02:30.765)
Yeah.

Kent C. Dodds (01:02:47.224)
Yeah.

Alex (01:02:54.085)
and you're going to start looking for patterns. You're going to say, okay, this a cluster of jargon showed up the most. These couple of worldviews keep showing up across every time I go out to do that. I'm to take a note of that. That's very interesting. Maybe I'll do some search specifically for that worldview and maybe use that as a filtering function for other bits of research and understanding. And you can see how these things kind of remix and fold into each other when you do that synthesis step intentionally and separately from the gathering step. And then where like the loop

Kent C. Dodds (01:03:10.179)
Hmm.

Kent C. Dodds (01:03:20.044)
Hmm.

Alex (01:03:22.055)
loops in on itself again is you take all that information and you plug it back into your search engine. You plug it back into Google. You can plug it back into your LLM of choice if you want to start exploring adjacencies. Like, all right, here's 10 terms of jargon that came up. What are 10 more that didn't come up that I can go plug into the same Discord or plug into the same Twitter list or whatever it is?

and see what comes to the surface. So each time you do this, you're both gathering insight as well as gathering fuel for what to look for and where to look for it next.

Kent C. Dodds (01:03:51.16)
Hmm.

Kent C. Dodds (01:03:58.711)
It sounds a little bit like you're saying the LLM can help you with gathering, but the synthesis needs to happen in your mind.

Alex (01:04:05.765)
I think that's very, very correct. I don't say that because I have a process that people pay money to learn. I think that is a very suspicious thing to come out of my mouth. I also say it because I've been trying it. I've gotten both, mean, Amy and I are objectively the best in the world at the process that we teach. And I've gotten pretty good at getting Claude and Claude Code in particular to do some truly remarkable things within the spectrum of this stuff.

Kent C. Dodds (01:04:14.973)
Kent C. Dodds (01:04:22.134)
Mm-hmm.

Alex (01:04:34.961)
But because I have the expertise, similar to the way a program has the expertise to know when an LLM is doing a thing that looks good but is actually a very bad idea, I have the expertise to look at it and go, there's stuff missing here. And it's not just like little stuff, it's like the core stuff. So yes, I think using it for the gathering, using it for organizing, using it to help you fill in gaps for yourself can be great. Again, I found these 10 terms.

Kent C. Dodds (01:04:47.361)
Kent C. Dodds (01:04:51.052)
Hmm.

Kent C. Dodds (01:05:00.782)
Mm.

Alex (01:05:04.295)
What are 10 others that I, like, in the past I'd tell people to do that manually and people would really struggle sometimes depending on their level of expertise. I don't think that's a bad thing because really what you're doing is you're basically, I don't know, I think about it as a bunch of like parallel paths or, I mean, we could think of this like threading almost, right? And it's basically like I'm spitting up a bunch of threads and each one has a new starting point. Some of them are ones that I picked.

There's no real risk of picking new starting points that the LLM helps you pick, so long as I go through the rest of the process to get it into my brain and I know what I'm looking for. I think making sure the LLM is your assistant, at best a collaborator, but I agree with you that the real magic here is in, if you're going to use an LLM, it's use the LLM to help you use your brain more for what it's good at, not less.

Kent C. Dodds (01:05:34.862)
Hmm.

Kent C. Dodds (01:05:54.446)
Hmm

Yeah, because ultimately the goal is to just really understand the space and have the words to communicate with people about the space, about the problem. And to understand also the competitors. You said like, what's being recommended and why? Like now you're being introduced to all the competitors and in that process of research, you can see where those competitors falls short. Maybe you see, I tried that guy, that product, but it didn't work for these reasons or whatever. Now you have like,

Alex (01:06:10.47)
Yeah?

Kent C. Dodds (01:06:25.646)
which is a really good understanding of the gaps in that space.

Alex (01:06:30.119)
I want to point out something you just said about, you're learning what terms to use. And I think that's a really good way to look at it when you're talking about communication is people listen more closely when they believe they are the ones being spoken to, right? And the terminology you choose is a good indicator that people respond to, not just consciously, but subconsciously. My mental model for this is when I'm learning these things and I want to practice them, I almost think about using other people's words, like putting on...

an outfit, like just trying on a new set of clothes. And like, I'm not getting married with clothes. I can take these clothes off anytime I want, but I want to see how, like I want to, I have to be able to say, I think you said this rather than the conversation too. I need to be able to not just say their words. I need to be able to say it and believe it. Right. It's like the believability is not a thing you can fake. If you're listening to this, you are not a good enough faker to get away with it. You just got to trust me on that. Like, like

Kent C. Dodds (01:07:03.885)
Heh.

Kent C. Dodds (01:07:16.534)
Hmm.

Kent C. Dodds (01:07:25.998)
you

Alex (01:07:28.997)
You have to be an extraordinary sociopath in order to really fake people out. It is super valuable to get those terms, get that language, get those ideas in your head, and you don't have to let it change you. But if you think about it like an outfit or a costume, if that's helpful, right? Say, I'm going to try on this particular perspective as a costume for a little bit and just participate in the conversation. It is remarkable how your brain and how quickly your brain will adapt to start thinking and feeling and responding.

Kent C. Dodds (01:07:40.43)
Hmm.

Alex (01:07:59.115)
more similarly to the people who you're trying to understand than not. doesn't change who you are and you can continue on being you, but it allows you to do almost like a version of code switching where, again, with practice, it stops feeling like I'm putting on a performance and starts feeling like this is just a natural mode of communication that I can intentionally choose when it is to our mutual benefit.

Kent C. Dodds (01:08:11.49)
Hmm.

Kent C. Dodds (01:08:26.722)
No, no, I love that. Alex, the only way that I could stop this conversation, because I feel like we have so much more to say, is that I have a hard stop in five minutes. Yeah.

Alex (01:08:34.439)
It's all good, all good. We can do a sequel in coming months if you like.

Kent C. Dodds (01:08:41.174)
I would love that. Alex, it's just been such a pleasure to have you on. I like to ask at the end of each episode a specific homework assignment that you can give people, something specific that they can do to improve their product sense, improve their identifying of audience, whatever, and something that they can put on a post-it note and then check it off when they're done. You got any homework for them?

Alex (01:09:04.903)
I love that you do this first of all. It's such a great way to wrap up a conversation like this. The thing that I'd love to encourage people to do is the next time you're in a conversation with your coworkers, your product team, whatever it is, and there is some form of disagreement around the direction, is to take a moment and try and step back and try to observe that conversation and ask the question, who is this?

Kent C. Dodds (01:09:07.212)
Hahaha

Alex (01:09:33.039)
disagreement in service of. And I like, it could be in service of the person with the positional power in the room. It could be the person, the decision maker could be the, I hope it's the client or the customer or the user. It could be the loudest person in the room, but just learning to observe who is this conversation in service of, I think is a really easy way that you can apply, like do it.

Kent C. Dodds (01:09:35.574)
Hmm.

Kent C. Dodds (01:09:47.702)
huh.

Alex (01:10:00.173)
in once a day, once a week, just as a practice. And to start noticing it and then start applying it to some of these other areas where there is an outcome that you're working towards. I think the reason for this is like you can observe without doing anything with the thing you observe. And that's why I kind of want to encourage it. And then as you start getting a sense of, I'm noticing some patterns here, then the upgrade of that exercise is how does your understanding of that influence what you might want to contribute to the conversation?

Kent C. Dodds (01:10:28.846)
Hmm.

Alex (01:10:29.947)
and then make that decision.

Kent C. Dodds (01:10:32.554)
I love it. Thank you, Alex. What's the best way for people to keep up with what you're doing and follow you on and like reach out for questions and things.

Alex (01:10:40.497)
Yeah, totally. So for all the product and stuff we talked about today, Stacking the Bricks, 30 by 500, stackingthebricks.com is the main website. I also have a book called the Tiny MBA. That's a tiny.mba. That's a Stacking the Bricks joint. And it's a fun, fast read that if this conversation was something you're into, I said, this is a business book for people who typically find business books itchy, like I said earlier. And I think the threads of

Kent C. Dodds (01:11:06.936)
Hmm.

Alex (01:11:10.215)
trust and values and consistency and service will resonate with your listeners. And I've gone back to being pretty active on Twitter again. It's Alex Hillman on Twitter. And these days it's more stuff around like cloud and cloud code because that has been like some of the most fun I've had exploring and learning something new. So if you were exploring that and learning stuff, give a follow over there and hit me up in the replies.

Kent C. Dodds (01:11:37.728)
Awesome. Hey, Alex, thank you so much for joining us and thanks everybody for listening and watching and we'll see you in the next one.