Jess Haberman: Hello and welcome to the Tech Podcast, a behind-the-scenes look at what it really takes to write and publish a technical book. I'm your host, Jess Heberman. I'm a literary agent, publishing consultant, and former acquisitions editor with nearly 20 years of experience publishing books. And today I'm thrilled to welcome Jacob Miller and Jeremy Mumford to the podcast. Hello to you both. ⁓ wanna share a brief intro, maybe starting with Jacob?
Jacob Miller: Yeah, so I'm the vice president of platform intelligence over AI at Pattern. It's an e-commerce technology accelerator and been here at Pattern about six years. Before that I've done academia, big corporations, and also an author now.
Jeremy Mumford: Yeah, my name is Jeremy Mumford. ⁓ I'm a lead AI engineer here at Pattern. I work closely with Jacob. ⁓ I've been at Pattern for four years. ⁓ before that, ⁓ I was at Brigham Young University where I was doing research and projects around deep fakes and AI image recognition.
Jess Haberman: Amazing. ⁓ well thank you both for joining today. ⁓ congratulations on publishing your first book, Architected Intelligence, Principles for Building AI First Organizations and Technologies. ⁓ so candidly, you guys and I worked together on the proposal. ⁓ we didn't work together in a traditional agent relationship, but it was really great to work with you on the proposal part of this project. ⁓ do you want to talk a little bit about sort of like what got you started on this whole journey? of what made you decide, hey, I want to write a book and and then how you know how it worked ⁓ in the in the proposal stage.
Jacob Miller: Yeah, I'll I'll dive in first. I I think it originated the idea of the book originated partly in presenting at conferences. There were ideas that we were presenting that really resonated with the audience that were more timeless in nature rather than just shifting with the technology. I think that was part of it. And second is we really saw a pain point where AI requires this multidisciplinary approach. And so different people would come from different backgrounds and make the mistakes that are missing. from their discipline that they needed to rely on others. And so that was the original concept as we were putting that together. So me and Jeremy were collaborating. This was what end of 2024, very end of 2024, when we were first going through this ideas of proposals. Also we ch I had some conversations with a a couple acquisition editors, I believe, who had just, you know, reaching out, they're always out talking and meeting people at conferences and such. And so that that started the process. And so we started putting together that outline, the concept of the book and the proposal. But our book was a little bit tricky because it's it's a hardcover book. It's not a traditional tech book with like here's blocks of code. We intentionally wrote the book so that it was frameworks and principles based, so that it would have been relevant three years ago and w it would be relevant five years from now, which is really important in AI because it's changing so quickly. And so we were focusing on those those foundational principles that like You're going to hit these problems even as the technology changes. So that's what made it kind of tricky as the concept of a book, because it's not purely tech, like here's these blocks of code type of book. And that was the that's where things got tricky as we were trying to navigate it. And then, you know, we we spoke with you, Jess, in terms of like how do we navigate this really awkward, kind of weird book that is doesn't quite fit the normal the normal sense.
Jess Haberman: 'Cause, you know, maybe talk a little bit about ⁓ you know, who were you writing for? You know, I think that that's probably where this helps the genre c sort of land more in like a business strategy or a t you know, tech strategy ⁓ sort of ⁓ topic area, right?
Jacob Miller: Yeah, do you wanna take that one, Jeremy?
Jeremy Mumford: Yeah, I think ⁓ for us when we were thinking about who the author and target audience would be, we were like, who who are we trying to reach? And I think for us it was tech leaders and tech architects, right? Now that can be anyone from the C suite level all the way down to senior engineers who are making architectural decisions. But thinking about how can we help those people who are charting these new territories of AI navigate. All of the complex decisions they have to make. There's so many unknowns, and so many people are jumping into disciplines that are potentially new to them. Software engineers have to now understand data science principles. Data scientists now have to understand software engineering principles. And leaders now need to fuse both of those disciplines together in this kind of new age of AI engineering. So our goal was to write a guide for those people.
Jess Haberman: So like how do you define like what is architected intelligence and and how does that connect to like the way we think about AI these days? Like AI's currently you know, really gone through a ⁓ an evolution in terms of how we define it. So ⁓ so what does architected intelligence mean in that case?
Jacob Miller: take this one. The ⁓ the architecture for us, it was solving AI problems, went through the same series of steps, basically in the same order, with the underlying principles and trade-offs for each one of those steps. And it started with the first one with output and saying, hey, is this truly valuable? How will it be used? How will it be used, you know, downstream, how will it be interpreted, and understanding that holistic experience of how you're executing. Because if it's not valuable, then then what are you doing? And especially right now in AI, And over the last couple years, one of our mantras has become interesting is only loosely correlated with valuable, because everyone would say, I can do this really interesting thing with AI. And so bringing people back and centering on how does that change the user experience, how does it change value. The second step was data, and data's changed a lot because unstructured data is now on the table. And unstructured data, you know, text and audio and all these bits of data that aren't in neatly defined you know rows and columns like they used to. And so the the management practices around that, the knowledge management systems that you need and those best practices around that is as it's changed and evolved. The third step is the models themselves, which most engineers are are more comfortable with. Things like context engineering and optimizing those those, optimizing your models, dealing with cost versus latency versus intelligence and managing that. Finally, that steps into observability, it's kind of this fourth step, and looking at it from a software engineer's perspective of understanding the logs, metrics, traces, like what is it doing? But then the data science observability concept of like, is it getting it right and why is it operating the way that it is? And the fifth kind of component, the fifth step of this is is with enablement in order to scale this. How do we do this not just as a one-off, but actually deliver this at scale and continually? And so whenever you're attacking an AI product or AI-infused process, you really need to hit each one of these. And it is a multidisciplinary approach. Jeremy mentioned, you know, data science and software engineering is the principle too, but there's also data engineering principles and knowledge management system principles that also weave throughout.
Jess Haberman: So mm ⁓ like what are the like the top ⁓ priority sort of problems that you're hoping your book is solving for these tech leaders? Like what are they struggling with the most right now and sort of where do you think this book can really help them?
Jeremy Mumford: I think for leaders making decisions, there are a few critical areas where they're being asked new questions that they haven't before. I think Jacob highlighted several of those. ⁓ I think there's a lot of questions around navigating unstructured data. ⁓ AI has kind of unlocked this whole new pool of data to be used at scale. ⁓ all these all this text and audio and images that previously you couldn't automate or process in a large scale way. Now you can with AI. So leaders are trying to figure out how can I take my extremely valuable data moat and combine it with AI in a way that's not just interesting but also useful. I think additionally there's questions around like what should an organization's North Star be, what output should they be targeting? ⁓ You see a lot of people just frantically trying to figure out how to use AI. You just quickly put a chat bot on their on their website, right? And you know Quickly putting a chatbot on your website to tell your customers like a quick summary of the article they're already looking at is not super useful. You know, how do we go deeper than just like slapping a chatbot on a website? And so that's like where we feel like our book really shines is really talking about how do we transform organizations in the same way that the internet era did or the electricity era did. A similar transformation needs to take place with AI that goes much deeper than just Some surface level changes.
Jess Haberman: Yeah, I hear you on kind of everyone's first step is like an AI rapper on whatever their their tech is. ⁓ so how do you like advise leaders to think about ⁓ like the ROI of AI or or efficiency gains or you know, cost savings? You know, like does it depend? Like ⁓ are there certain maybe met like leading metrics that that ⁓ folks maybe are missing because
Jeremy Mumford: Yeah. Yeah.
Jess Haberman: there may be so focused on ⁓ like you know r maybe reducing headcount or you know being as efficient as possible in terms of the organization or making more money or you know like what are the what are the key areas where you think these leaders are like maybe ⁓ should focus and maybe aren't really maybe it's a blind spot for them right now.
Jacob Miller: Yeah, one of the one of the principles that we emphasize early on in the book in that output section of what's valuable is that AI is worth applying to the core of whatever it is that you're offering. And so if I'm in no matter what industry I'm in, what is what defines my company, what defines my organization, why do people find what I provide valuable? And that is what I need to be applying AI to. In order to drive the existing metrics that already exist. And if I can't figure that out, one of my competitors will, and that is an existential question. And so the the concept of, hey. AI is really cool. I'm gonna apply it to all the things that I don't like doing and aren't central to my company, just so I can trim some headcount around the edges or something like that for efficiency. It's one of the messages that we try to pull people away from and say like, no, why do you exist? What do you need to effectively drive? And then work backward from there. And then there I mean there there's all sorts of proxy metrics depending on the industry that you're in, in terms of the efficiency and the effectiveness that you're getting out of those applications and models. But it's so easy to take your eye off that North Star of like, why do we even exist as a company or as an organization? And that's what we should deliver first.
Jeremy Mumford: Yeah. And just to like tap on something to what Jacob said, ⁓ Jensen Huang, the the CEO of NVIDIA said, if your best idea for increasing ROI with AI is laying people off, you lack imagination. Something to that effect. And so hopefully people have a little more imagination than just let's cut head count. Ideally, if you leverage AI correctly, you should actually be hiring people 'cause your company is growing so fast. So
Jess Haberman: Yeah. Yeah. I like that, you know, your your business exists for a reason. There is some core competitive advantage that you can maximize rather than, yeah, like kind of focus on the edges. But yeah, how do you how do you apply AI, you know, smartly in in and to make yourselves like stand out even more and what you already do well. So ⁓ I think that's a message leaders and developers and data scientists like all across the board can get behind for sure. that's great. ⁓ you know And how much do you how much do you get into kind of your the sort of technology architecture of your of your your tech stack, your you know, the way that maybe you're ⁓ working with models? Like how do you how far into the sort of technical tinkering of your infrastructure do you do you advise in your book?
Jeremy Mumford: I can take this one. it's a fun question of like how deep on the technical do we go? like this is something that like Jacob would have to pull me back on, is because I like to go deep on the technical side. And I think the North Star for our book was, as Jacob said earlier, let's make sure that this book ⁓ is enduring. We wanted to make sure that five years from now, ten years from now, this book is still useful to people. And so we made a very conscious decision at the beginning that ⁓ we would explicitly never name a specific tool, a specific vendor or provider, a specific AI model. We would never include any code throughout the book. ⁓ and that decision we feel has paid dividends ⁓ because as you read through the book, it all feels extremely relevant. ⁓ this is a common thing I find when reading AI books. I go through and I read it, and they say use. This code with this model and instantly I'm like, okay, well that's outdated as of like, you know, a month after release. And so ⁓ we're really excited about how we intentionally structured the book in that way by being very descriptive and detailed in how people should architecture their approaches and decision making and design in a way that technical leaders will benefit from, but without explicitly naming any tools or vendors or code that we feel will be outdated.
Jacob Miller: Yeah. And I would add to that, we were still able to provide the nuance and depth. And to the degree that we would provide a specific example, say like temperature within an AI model, we would say here's the underlying principle. Of how to create this variation in approaches. And some of the current examples that exist today are to modify things like A, B, and C. So we would, to the extent that we would center it on today's technology, it would only be as examples rather than this is how you create the variation. We say, no, no, no, here's the pr underlying principle and here's an example expression of it. Whatever that however the technology changes, that principle will endure.
Jess Haberman: Yeah, I think that's really smart in terms of kind of trying to keep your book evergreen as long as possible. It's it's hard these days to keep a book like you know, relevant for three to five years. That seems like ⁓ so insane where it that may have been easier in the past. But ⁓ but yeah, no, that makes a lot of sense. ⁓ let's talk a little bit about ⁓ your process. You know, I we were not you know, I w I wasn't communicating with you quite a lot during the process, so I'm I would love to hear more about, you know, what did you what did you learn? What did you like wish that you knew earlier that you might know now after having been through that sort of end-to-end process of publishing a book?
Jacob Miller: I I can start with a few thoughts and ideas, and Jeremy has a ton to add as well as we both navigated this experience. First, it was it was difficult because we had an awkward audience that we were shooting for. I mean, we're it's like a hardcover book, it's a little bit different, but it's tech, but it's not code in tech. So it's it's this it was this weird niche. One of the things that I didn't understand early on is that you have these publishers and some of them we we'd been talking to. Some of them we wanted to reach out to is that they have all of these forms of like, hey, this is what we want like a book proposal to look like. And we did ever we didn't I didn't know until you know working with you, Jess, that those were guidelines rather than strict, this is exactly how your book proposal should look like. And talking to some of the publishers, it was interesting how they said, ⁓ I really like this part of your book proposal. And It was something that they didn't ask for as part of like their standard book proposal form. And so the idea of they do have like a book proposal form that's kind of a format that they're looking for, but really we want to we don't want that to constrain how we're approaching the situation. We want to put our best foot forward because we're trying to make it easy and simple for them to understand who our audience is, how we're approaching it, and what those ideas
Jess Haberman: Yeah. It's helpful because publishers don't look at each other's proposal guidelines. You know, they don't they're not like they have no idea what, you know, O'Reilly might be asking for one thing, Wiley's asking for something else, Pierce is asking for another, and they're not paying attention necessarily. Why like why would they? But a good acquisitions editor is gonna find like the the the sort of gold nugget in in whatever template it's in. ⁓ so yeah, that's a good one. I was ⁓ I think you made a good point about your audience being a little awkward, being on a niche in the and it being tough to really narrow to that focus. But I think that's a huge selling point for for anyone thinking about writing a book that you if you I think one of the biggest misconceptions in publishing is I've got to write a book for for a big audience. ⁓ like no if you can speak directly to a like a specific audience you don't want a very small one but tech leaders is still a big audience you know and and and sort of decision makers and tech Tech companies. And like you said, you know, that also includes like, you know, you know, senior architects and things like that. So ⁓ so if you know exactly who you're speaking to, and you know, your everything you write is really tailored to that audience, it can work really, really well. It's I see folks struggle when they're trying to also talk to developers while also talking to leaders. Like it's so hard to hit the same message for you know, for people who have really different goals and motivations. ⁓ So what do you think like helped you like really speak to that audience? Did you interview folks? You know, you work with tech leaders all the time, so I imagine a lot of your own experience really helped you there, but how did you like hone kind of hone your ⁓ your story, your narrative for the audience that you think it would resonate with best?
Jeremy Mumford: Yeah, I think it was a combination of two things, understanding ⁓ who to write the book. Them the most. I think, first off, Jacob and myself have a network of people that we talk to on a regular basis. And we started to see a pattern of common questions and common needs across the industry. But number two, all of those things tied back to things we were seeing ourselves in our current job and our current roles of the company. Combining that broad overall view we saw across the industry with our personal experiences, we felt we had a really strong vision for who needed this book and what it needed to be about.
Jess Haberman: Jeremy, did you have any particular like major learnings from the writing process itself? Like any like things that you sort of ⁓ maybe was a misconception early on and that you sort of figured out as you went along?
Jeremy Mumford: Are you saying a misconception about the subject material or a misconception about book writing?
Jess Haberman: ⁓ well I meant book writing, but either way, like I think both are interesting.
Jeremy Mumford: 'Cause I mean I think that happens too. Anytime you like spend a significant amount of time discussing a subject material, I think you discover new things about it, overturn previous conception ⁓ ideas and concepts. ⁓ but as far as book writing specifically, ⁓ it was very eye-opening for me. I think Jacob had written a couple of textbooks before and been involved with some writing in that way. ⁓ the extent of my writing had been college-based. I you know, I did a thesis and I did papers in college, of course. ⁓ but writing something to the level of a a whole book with a with a traditional publisher is a very large project. And I don't think I really knew what I was getting into when I agreed to sign up with Jacob, how long and big of a project this would be. Well worth it. but I had no idea. I had no idea of all the different phases of development, how you get assigned a development editor and then a copy editor and then a layout editor.
Jess Haberman: What?
Jeremy Mumford: I didn't realize like the complex communication structure between different representatives at the publisher. I didn't realize that the proposal process could be tackled in such like a vigorous, intense way and that there are certain ways you need to write the proposal and contact certain people or you'll just disappear into the pile. ⁓ so all of these things were a learning experience. And then also just learning about how to build the habits of of writing, you know, all all those sort of habits that come in and make it so you actually Make your deadline. ⁓
Jess Haberman: Yeah, right. So how did you guys manage it as co-authors? I've seen this work really well in some cases and I've seen it be a complete disaster in other cases. So, you know, it sounds like you guys are still friends. That's great. Good to see that, you know. So you survived writing a book together. So how did how did it work for you? Where did you kind of find routines that made it, you know, made it work? And where like where was their tension?
Jeremy Mumford: Yeah, I think ⁓ I'll take this one again if that's okay, Jacob. I think for us it was finding time each week to visit each other's houses and just sit down and write the book. What would often happen is ⁓ Jacob would come over to my place one or two times a week and we would cast Google Docs up on the TV and we would vigorously debate a concept or an idea and Just kind of imagine the two of us just pacing around the living room, gesturing at the screen, getting really excited about ideas. And then we'd maybe pause and we'd like each silently write for thirty minutes. And then we'd reconvene and we'd kind of do it all over again. And so there's a lot of rounds of that. And then often what it is is like one person would write a chapter and then they would pass it off to the other person and that person would go through the entire chapter and kind of put their own polish and spin on it. And then we'd go back to the other person and that person would do that process again. And we did that for all 12 chapters. So very time consuming, but I feel like the end result became extremely polished as a result.
Jacob Miller: I will say it's a lot easier to go if I if I weren't co-authoring it, I think A, the quality would have been worse because like having another person to just be like, Are you sure about that? Just someone else to prod that and like that's kind of a bad idea. which Jeremy does in an excellent diplomatic way, but he's very good at communicating, like, are you sure that's what we want to communicate on one part. And then the second is that there were certain areas of the book where each of us were bringing expertise. So certain chapters, it was like, ⁓ Jeremy's leading on this in chapter nine, because he just knows it better than I do. But I'm over here in chapter five and I know chapter five and I know that content better than Jeremy does. That doesn't mean Jeremy doesn't contribute to chapter five, but there's kind of like a designated lead for that chapter because it follows their own subject matter expertise. So I I think having that as a natural lead proportion and then going back and forth was a a super healthy dynamic. I mean we're we're still friends. We still I think we're more friends now than we were before even. And so we succeeded, I think.
Jeremy Mumford: Okay.
Jess Haberman: Present. That's great. Yeah, I think ⁓ collaboration is really important even when you are the s a sole author, ⁓ to have tech reviewers who can really kind of push against some of your ⁓ assumptions and can can make your help you make your points better ⁓ is is really important. ⁓ but that's great that you guys had kind of a like whiteboarding slash writing retreat ⁓ kind of structure. That sounds a lot of fun. ⁓ I like that. I I like that a lot. and then I think we all know ⁓ publish. one book doesn't make you rich overnight. ⁓ you know, but Jeremy, you mentioned that, you know, it it was worth it, it it's still valuable. ⁓ tell, you know, what do you what are your I know it's early on in the in the process of of the you know publication, but what are your kind of expectations going in and ⁓ what are you hoping that this book will enable for you both in terms of your careers and your and and and future work and things like that.
Jeremy Mumford: Yeah. I mean it should hopefully come as no surprise to anybody that writing a book is not a get rich quick scheme. you know. But I think ⁓ we went into it, you know, eyes wide open with, you know, the purpose of us writing this book is ⁓ I think one, Jacob shared that he had some ideas that he was very passionate about, ⁓ and he wanted to share those with the world. And I think that should really be like the root reason for anybody wanting to write a book is you have something you want to share. With a part of the world. ⁓ I think two, I think this will benefit us moving forward. I think there will be benefits to our personal knowledge and understanding of AI. I feel like both Jacob and myself have really honed our understanding of how AI works and at an organizational level to such a degree, ⁓ because of writing the book. And then of course, there's all the benefits that come from being able to say that you wrote a book on a subject like this. There there are benefits that come from that. ⁓ but I think for me it's it was really the process of writing the book and the knowledge gained from it. That is probably the bigger factor for me. And then quite selfishly the sense of pride of being able to say I'm an author is very fun as well. Being able to tell people, ⁓ yeah, I wrote a book. That's pretty cool. You know.
Jacob Miller: I'll add that it does require a certain level of hubris to in art or as an author or anything like that because there are a million, you know, content creators out there. And especially if you're focusing on something deep, you you know, you're not just throwing out some 45 second TikTok. Like this is something that's going to take like hundreds of hours of your time that you're going to dedicate to this. And it takes a certain level of like hubris to say, like, well, I have something to say and the world doesn't have it otherwise. And to be able to push that forward. But that's where that passion comes from, because you know, the world I love the quote, the world is a museum of passion projects. The world is full of these passion projects that people have pursued. And I don't think it is accept you know, doesn't mean you're exceptionally full of like hubris. If you're writing something that you're passionate about, because you're bringing that passion, that depth of expertise to share with everybody. And I would add that the process does force clarity, like that discipline and clarity. And so I I love the line, I think it was Paul Graham who's popularized it, but it's to write is to think is writing forces you to clarify your thoughts and put them on paper and adjust go through it slowly and carefully. I think that's one of the drawbacks of AI writing, actually, is like, ⁓ I had some general concepts, then AI filled it out. the pressure process of like AI is tremendously hu useful. It was very useful for us to help us with the writing, but you going through the process and pain of writing and the trade-offs and what we truly mean by every word is something that sharpens your understanding of any field.
Jess Haberman: Yeah, I love that. Yeah, I I do think ⁓ publishing a book does require a little bit of hubris a lot of curiosity and patience. ⁓ so yeah, it's that that's a great lesson. you know, are there any things that for you know, for other people out there who are interested in writing a a technical or a business book, ⁓ are there any things that you would warn them, ⁓ and now that you've been through the process of of writing your own?
Jacob Miller: I would start with it's going to be take a lot of time. If you're not willing to dedicate the time and have the passion behind it. Like if you're not to Jeremy's point from earlier, if you don't have the passion behind it, you might not make it through. Because if you know, on a per hour basis of like how much money do you make out of this, it's probably not some astronomical figure compared to what you could be doing for your day job or consulting or doing something else. It's really driven by this underlying passion and drive. I would say that's one part
Jeremy Mumford: Yeah, I think heads up on the time, like buckle up, you're you're signing for a lot of hours. I think that's probably the biggest gotcha. ⁓
Jacob Miller: I I would say as first time authors having your your help, Jess, throughout the process for as a first time author, that's extremely helpful. probably like second or third or fourth, maybe like they've kind of figured it out at that point and kinda know the ropes and know their way through it. But it is pretty difficult as first time authors 'cause there's there's so much in the domain that like even if you ask AI to help you do something, it's not navigating like those relationships and the expectations and what are the you know, what are the incentives and objectives of the people on the other side of the table. It's really hard to understand what that is without some type of assistance and help.
Jess Haberman: It is interesting that publishing is centuries old and ⁓ AI still isn't trained on enough material to help you do do it step by step. It's it's such a unique process and it is so based on relationships and you know how you navigate them. So that's really interesting. so and and then I I'm curious kind of like what do you what do you guys read? What kind of ⁓ content do you take in as your ⁓ you know, what informed your your writing and ⁓ did you were you inspired by other books or other pieces of media ⁓ that sort of inspired the work that you did?
Jeremy Mumford: Yeah, I think as far as reading, ⁓ for me there's kind of two sides to my reading life. There's like books and content I read for fun. ⁓ that's a lot of like Brandon Sanderson and like Golden Age Science Fiction. ⁓ a whole range of books, you know, I I absolutely love reading. ⁓ then on on the more technical side, ⁓ I do like to get into technical books and we had several that we we looked to for examples. ⁓ You know, there's books like Observability that wa that was really fantastic, ⁓ and a few others. LM can it's the LM Engineers Handbook was another favorite of mine. ⁓ I think that one's by Paul Utsin. ⁓ that was kind of a helpful example of writing for AI in the modern age. ⁓ and we would reach out to these authors on on LinkedIn and other platforms and ask them for advice and say, What would you advise about writing? And so, you know, don't be afraid to Hit people up. it's a very friendly community. People love to share ideas. You know, we even got some local Utah authors that we talked to and asked for advice on which publishers they preferred and ⁓ what writing techniques and styles they recommend and habits they would recommend for us. So all of that was very helpful.
Jacob Miller: Yeah, Jeremy highlighted, for example, observability was one book. There were certain books that we saw within the text space that were at least a few years old that were still hyper relevant over time. And especially certain chapters of them that were hyper relevant. And so like you'd you'd read this chapter, you know, within a larger book, and like you could have read it fifty years ago or ten years ago or two years ago or today or five years from now. And it's like it still resonates, it's still true. And that was like That's w exactly what we were shooting for, but with the audience that we had in mind and then on the topic that we cared about. So combining those three, there were these lots of like these specific examples, and they usually were like a few chapters within a larger book where we'd you know, we'd read those books and they really stuck with us over many, many years.
Jess Haberman: So ⁓ so what's next for you? Are you thinking about writing another book after this? So you need a break or ⁓ are you ready ready to keep keep going?
Jeremy Mumford: I mean, I need a break. It's that was a lot of work. I think next steps for us, obviously, right now at the time of recording, the book is still under pre order by the time this comes out. It'll be available on Amazon and Barnes and Nobles. And I think we're excited to go out and share it with the world after having worked on it so hard over the last year. We're really excited for people to dive into it, see what they think. ⁓ And then we're busy kind of applying all of those concepts we talked about at pattern. So, you know, we are practitioners of what we do right now. We're we're very busy building a lot of stuff. ⁓ and that's a lot of fun too. So that's kind of the immediate future for for both of us.
Jacob Miller: Yeah, we've got a lot of building, some consulting and some helping as part of that, but also some conferences, stuff like podcasts to share those the key learnings and key concepts and ideas. ⁓ there's there has been a book that I've been like working on kind of within the AI space, but as a little bit more general audience as far as the lessons that you need to know in interacting with AI systems as like regular people, with the more general audience, but that's someday maybe that will see the light of day. For now it's it's essays that I post on on the blog and things like that.
Jess Haberman: Yeah. That's brilliant. ⁓ I mean it it it is impressive that you published this whole book and had a full-time job in the meantime. ⁓ so congrats on that. Major kudos. ⁓ yeah, and I think like I tell authors all the time that get you know, getting the book published is a huge feat and it's worth celebrating, but it's just step one. It's there's you know, you gotta get the word out. You've invested so much time and energy in writing and then, you know, it's it does require still a lot more energy to get it in front of people. So ⁓ so don't be shy. definitely make sure everybody knows about it. Hit them over the head with it. ⁓ because it's really important. It's so timely too. I think it's a great topic that everyone's already talking about. ⁓ So yeah. Is there anything else? I you Jacob, you mentioned your ⁓ your blog, but is there anything else you want to plug? Where can folks find out more about you?
Jacob Miller: Yeah, so we write at the whybehindai dot com. It's where we post essays, we keep things a little bit lighter there. Many concepts from the from our book end up there or have ended up there, although the blog form is a little bit lighter and more entertaining and has better memes, to be honest. And then the book itself is Architected Intelligence, and you can find us on LinkedIn, Jacob Miller and Jeremy Mumford. Did I miss anything, Jeremy?
Jess Haberman: Amazing.
Jeremy Mumford: I think that's it.
Jess Haberman: Awesome. Well, thank you both for joining the pod today. It's been a pleasure speaking to you and hearing more about the book and and the whole process you went through. Congrats again. ⁓ please everyone do order a copy of Architected Intelligence available where books are sold. And ⁓ hope to talk to you guys again soon. I can't wait to see where this goes, and I hope all the success for your book.
Jacob Miller: Thanks, Jess.
Jeremy Mumford: Thank you.
Jacob Miller: Yeah, so I'm the vice president of platform intelligence over AI at Pattern. It's an e-commerce technology accelerator and been here at Pattern about six years. Before that I've done academia, big corporations, and also an author now.
Jeremy Mumford: Yeah, my name is Jeremy Mumford. ⁓ I'm a lead AI engineer here at Pattern. I work closely with Jacob. ⁓ I've been at Pattern for four years. ⁓ before that, ⁓ I was at Brigham Young University where I was doing research and projects around deep fakes and AI image recognition.
Jess Haberman: Amazing. ⁓ well thank you both for joining today. ⁓ congratulations on publishing your first book, Architected Intelligence, Principles for Building AI First Organizations and Technologies. ⁓ so candidly, you guys and I worked together on the proposal. ⁓ we didn't work together in a traditional agent relationship, but it was really great to work with you on the proposal part of this project. ⁓ do you want to talk a little bit about sort of like what got you started on this whole journey? of what made you decide, hey, I want to write a book and and then how you know how it worked ⁓ in the in the proposal stage.
Jacob Miller: Yeah, I'll I'll dive in first. I I think it originated the idea of the book originated partly in presenting at conferences. There were ideas that we were presenting that really resonated with the audience that were more timeless in nature rather than just shifting with the technology. I think that was part of it. And second is we really saw a pain point where AI requires this multidisciplinary approach. And so different people would come from different backgrounds and make the mistakes that are missing. from their discipline that they needed to rely on others. And so that was the original concept as we were putting that together. So me and Jeremy were collaborating. This was what end of 2024, very end of 2024, when we were first going through this ideas of proposals. Also we ch I had some conversations with a a couple acquisition editors, I believe, who had just, you know, reaching out, they're always out talking and meeting people at conferences and such. And so that that started the process. And so we started putting together that outline, the concept of the book and the proposal. But our book was a little bit tricky because it's it's a hardcover book. It's not a traditional tech book with like here's blocks of code. We intentionally wrote the book so that it was frameworks and principles based, so that it would have been relevant three years ago and w it would be relevant five years from now, which is really important in AI because it's changing so quickly. And so we were focusing on those those foundational principles that like You're going to hit these problems even as the technology changes. So that's what made it kind of tricky as the concept of a book, because it's not purely tech, like here's these blocks of code type of book. And that was the that's where things got tricky as we were trying to navigate it. And then, you know, we we spoke with you, Jess, in terms of like how do we navigate this really awkward, kind of weird book that is doesn't quite fit the normal the normal sense.
Jess Haberman: 'Cause, you know, maybe talk a little bit about ⁓ you know, who were you writing for? You know, I think that that's probably where this helps the genre c sort of land more in like a business strategy or a t you know, tech strategy ⁓ sort of ⁓ topic area, right?
Jacob Miller: Yeah, do you wanna take that one, Jeremy?
Jeremy Mumford: Yeah, I think ⁓ for us when we were thinking about who the author and target audience would be, we were like, who who are we trying to reach? And I think for us it was tech leaders and tech architects, right? Now that can be anyone from the C suite level all the way down to senior engineers who are making architectural decisions. But thinking about how can we help those people who are charting these new territories of AI navigate. All of the complex decisions they have to make. There's so many unknowns, and so many people are jumping into disciplines that are potentially new to them. Software engineers have to now understand data science principles. Data scientists now have to understand software engineering principles. And leaders now need to fuse both of those disciplines together in this kind of new age of AI engineering. So our goal was to write a guide for those people.
Jess Haberman: So like how do you define like what is architected intelligence and and how does that connect to like the way we think about AI these days? Like AI's currently you know, really gone through a ⁓ an evolution in terms of how we define it. So ⁓ so what does architected intelligence mean in that case?
Jacob Miller: take this one. The ⁓ the architecture for us, it was solving AI problems, went through the same series of steps, basically in the same order, with the underlying principles and trade-offs for each one of those steps. And it started with the first one with output and saying, hey, is this truly valuable? How will it be used? How will it be used, you know, downstream, how will it be interpreted, and understanding that holistic experience of how you're executing. Because if it's not valuable, then then what are you doing? And especially right now in AI, And over the last couple years, one of our mantras has become interesting is only loosely correlated with valuable, because everyone would say, I can do this really interesting thing with AI. And so bringing people back and centering on how does that change the user experience, how does it change value. The second step was data, and data's changed a lot because unstructured data is now on the table. And unstructured data, you know, text and audio and all these bits of data that aren't in neatly defined you know rows and columns like they used to. And so the the management practices around that, the knowledge management systems that you need and those best practices around that is as it's changed and evolved. The third step is the models themselves, which most engineers are are more comfortable with. Things like context engineering and optimizing those those, optimizing your models, dealing with cost versus latency versus intelligence and managing that. Finally, that steps into observability, it's kind of this fourth step, and looking at it from a software engineer's perspective of understanding the logs, metrics, traces, like what is it doing? But then the data science observability concept of like, is it getting it right and why is it operating the way that it is? And the fifth kind of component, the fifth step of this is is with enablement in order to scale this. How do we do this not just as a one-off, but actually deliver this at scale and continually? And so whenever you're attacking an AI product or AI-infused process, you really need to hit each one of these. And it is a multidisciplinary approach. Jeremy mentioned, you know, data science and software engineering is the principle too, but there's also data engineering principles and knowledge management system principles that also weave throughout.
Jess Haberman: So mm ⁓ like what are the like the top ⁓ priority sort of problems that you're hoping your book is solving for these tech leaders? Like what are they struggling with the most right now and sort of where do you think this book can really help them?
Jeremy Mumford: I think for leaders making decisions, there are a few critical areas where they're being asked new questions that they haven't before. I think Jacob highlighted several of those. ⁓ I think there's a lot of questions around navigating unstructured data. ⁓ AI has kind of unlocked this whole new pool of data to be used at scale. ⁓ all these all this text and audio and images that previously you couldn't automate or process in a large scale way. Now you can with AI. So leaders are trying to figure out how can I take my extremely valuable data moat and combine it with AI in a way that's not just interesting but also useful. I think additionally there's questions around like what should an organization's North Star be, what output should they be targeting? ⁓ You see a lot of people just frantically trying to figure out how to use AI. You just quickly put a chat bot on their on their website, right? And you know Quickly putting a chatbot on your website to tell your customers like a quick summary of the article they're already looking at is not super useful. You know, how do we go deeper than just like slapping a chatbot on a website? And so that's like where we feel like our book really shines is really talking about how do we transform organizations in the same way that the internet era did or the electricity era did. A similar transformation needs to take place with AI that goes much deeper than just Some surface level changes.
Jess Haberman: Yeah, I hear you on kind of everyone's first step is like an AI rapper on whatever their their tech is. ⁓ so how do you like advise leaders to think about ⁓ like the ROI of AI or or efficiency gains or you know, cost savings? You know, like does it depend? Like ⁓ are there certain maybe met like leading metrics that that ⁓ folks maybe are missing because
Jeremy Mumford: Yeah. Yeah.
Jess Haberman: there may be so focused on ⁓ like you know r maybe reducing headcount or you know being as efficient as possible in terms of the organization or making more money or you know like what are the what are the key areas where you think these leaders are like maybe ⁓ should focus and maybe aren't really maybe it's a blind spot for them right now.
Jacob Miller: Yeah, one of the one of the principles that we emphasize early on in the book in that output section of what's valuable is that AI is worth applying to the core of whatever it is that you're offering. And so if I'm in no matter what industry I'm in, what is what defines my company, what defines my organization, why do people find what I provide valuable? And that is what I need to be applying AI to. In order to drive the existing metrics that already exist. And if I can't figure that out, one of my competitors will, and that is an existential question. And so the the concept of, hey. AI is really cool. I'm gonna apply it to all the things that I don't like doing and aren't central to my company, just so I can trim some headcount around the edges or something like that for efficiency. It's one of the messages that we try to pull people away from and say like, no, why do you exist? What do you need to effectively drive? And then work backward from there. And then there I mean there there's all sorts of proxy metrics depending on the industry that you're in, in terms of the efficiency and the effectiveness that you're getting out of those applications and models. But it's so easy to take your eye off that North Star of like, why do we even exist as a company or as an organization? And that's what we should deliver first.
Jeremy Mumford: Yeah. And just to like tap on something to what Jacob said, ⁓ Jensen Huang, the the CEO of NVIDIA said, if your best idea for increasing ROI with AI is laying people off, you lack imagination. Something to that effect. And so hopefully people have a little more imagination than just let's cut head count. Ideally, if you leverage AI correctly, you should actually be hiring people 'cause your company is growing so fast. So
Jess Haberman: Yeah. Yeah. I like that, you know, your your business exists for a reason. There is some core competitive advantage that you can maximize rather than, yeah, like kind of focus on the edges. But yeah, how do you how do you apply AI, you know, smartly in in and to make yourselves like stand out even more and what you already do well. So ⁓ I think that's a message leaders and developers and data scientists like all across the board can get behind for sure. that's great. ⁓ you know And how much do you how much do you get into kind of your the sort of technology architecture of your of your your tech stack, your you know, the way that maybe you're ⁓ working with models? Like how do you how far into the sort of technical tinkering of your infrastructure do you do you advise in your book?
Jeremy Mumford: I can take this one. it's a fun question of like how deep on the technical do we go? like this is something that like Jacob would have to pull me back on, is because I like to go deep on the technical side. And I think the North Star for our book was, as Jacob said earlier, let's make sure that this book ⁓ is enduring. We wanted to make sure that five years from now, ten years from now, this book is still useful to people. And so we made a very conscious decision at the beginning that ⁓ we would explicitly never name a specific tool, a specific vendor or provider, a specific AI model. We would never include any code throughout the book. ⁓ and that decision we feel has paid dividends ⁓ because as you read through the book, it all feels extremely relevant. ⁓ this is a common thing I find when reading AI books. I go through and I read it, and they say use. This code with this model and instantly I'm like, okay, well that's outdated as of like, you know, a month after release. And so ⁓ we're really excited about how we intentionally structured the book in that way by being very descriptive and detailed in how people should architecture their approaches and decision making and design in a way that technical leaders will benefit from, but without explicitly naming any tools or vendors or code that we feel will be outdated.
Jacob Miller: Yeah. And I would add to that, we were still able to provide the nuance and depth. And to the degree that we would provide a specific example, say like temperature within an AI model, we would say here's the underlying principle. Of how to create this variation in approaches. And some of the current examples that exist today are to modify things like A, B, and C. So we would, to the extent that we would center it on today's technology, it would only be as examples rather than this is how you create the variation. We say, no, no, no, here's the pr underlying principle and here's an example expression of it. Whatever that however the technology changes, that principle will endure.
Jess Haberman: Yeah, I think that's really smart in terms of kind of trying to keep your book evergreen as long as possible. It's it's hard these days to keep a book like you know, relevant for three to five years. That seems like ⁓ so insane where it that may have been easier in the past. But ⁓ but yeah, no, that makes a lot of sense. ⁓ let's talk a little bit about ⁓ your process. You know, I we were not you know, I w I wasn't communicating with you quite a lot during the process, so I'm I would love to hear more about, you know, what did you what did you learn? What did you like wish that you knew earlier that you might know now after having been through that sort of end-to-end process of publishing a book?
Jacob Miller: I I can start with a few thoughts and ideas, and Jeremy has a ton to add as well as we both navigated this experience. First, it was it was difficult because we had an awkward audience that we were shooting for. I mean, we're it's like a hardcover book, it's a little bit different, but it's tech, but it's not code in tech. So it's it's this it was this weird niche. One of the things that I didn't understand early on is that you have these publishers and some of them we we'd been talking to. Some of them we wanted to reach out to is that they have all of these forms of like, hey, this is what we want like a book proposal to look like. And we did ever we didn't I didn't know until you know working with you, Jess, that those were guidelines rather than strict, this is exactly how your book proposal should look like. And talking to some of the publishers, it was interesting how they said, ⁓ I really like this part of your book proposal. And It was something that they didn't ask for as part of like their standard book proposal form. And so the idea of they do have like a book proposal form that's kind of a format that they're looking for, but really we want to we don't want that to constrain how we're approaching the situation. We want to put our best foot forward because we're trying to make it easy and simple for them to understand who our audience is, how we're approaching it, and what those ideas
Jess Haberman: Yeah. It's helpful because publishers don't look at each other's proposal guidelines. You know, they don't they're not like they have no idea what, you know, O'Reilly might be asking for one thing, Wiley's asking for something else, Pierce is asking for another, and they're not paying attention necessarily. Why like why would they? But a good acquisitions editor is gonna find like the the the sort of gold nugget in in whatever template it's in. ⁓ so yeah, that's a good one. I was ⁓ I think you made a good point about your audience being a little awkward, being on a niche in the and it being tough to really narrow to that focus. But I think that's a huge selling point for for anyone thinking about writing a book that you if you I think one of the biggest misconceptions in publishing is I've got to write a book for for a big audience. ⁓ like no if you can speak directly to a like a specific audience you don't want a very small one but tech leaders is still a big audience you know and and and sort of decision makers and tech Tech companies. And like you said, you know, that also includes like, you know, you know, senior architects and things like that. So ⁓ so if you know exactly who you're speaking to, and you know, your everything you write is really tailored to that audience, it can work really, really well. It's I see folks struggle when they're trying to also talk to developers while also talking to leaders. Like it's so hard to hit the same message for you know, for people who have really different goals and motivations. ⁓ So what do you think like helped you like really speak to that audience? Did you interview folks? You know, you work with tech leaders all the time, so I imagine a lot of your own experience really helped you there, but how did you like hone kind of hone your ⁓ your story, your narrative for the audience that you think it would resonate with best?
Jeremy Mumford: Yeah, I think it was a combination of two things, understanding ⁓ who to write the book. Them the most. I think, first off, Jacob and myself have a network of people that we talk to on a regular basis. And we started to see a pattern of common questions and common needs across the industry. But number two, all of those things tied back to things we were seeing ourselves in our current job and our current roles of the company. Combining that broad overall view we saw across the industry with our personal experiences, we felt we had a really strong vision for who needed this book and what it needed to be about.
Jess Haberman: Jeremy, did you have any particular like major learnings from the writing process itself? Like any like things that you sort of ⁓ maybe was a misconception early on and that you sort of figured out as you went along?
Jeremy Mumford: Are you saying a misconception about the subject material or a misconception about book writing?
Jess Haberman: ⁓ well I meant book writing, but either way, like I think both are interesting.
Jeremy Mumford: 'Cause I mean I think that happens too. Anytime you like spend a significant amount of time discussing a subject material, I think you discover new things about it, overturn previous conception ⁓ ideas and concepts. ⁓ but as far as book writing specifically, ⁓ it was very eye-opening for me. I think Jacob had written a couple of textbooks before and been involved with some writing in that way. ⁓ the extent of my writing had been college-based. I you know, I did a thesis and I did papers in college, of course. ⁓ but writing something to the level of a a whole book with a with a traditional publisher is a very large project. And I don't think I really knew what I was getting into when I agreed to sign up with Jacob, how long and big of a project this would be. Well worth it. but I had no idea. I had no idea of all the different phases of development, how you get assigned a development editor and then a copy editor and then a layout editor.
Jess Haberman: What?
Jeremy Mumford: I didn't realize like the complex communication structure between different representatives at the publisher. I didn't realize that the proposal process could be tackled in such like a vigorous, intense way and that there are certain ways you need to write the proposal and contact certain people or you'll just disappear into the pile. ⁓ so all of these things were a learning experience. And then also just learning about how to build the habits of of writing, you know, all all those sort of habits that come in and make it so you actually Make your deadline. ⁓
Jess Haberman: Yeah, right. So how did you guys manage it as co-authors? I've seen this work really well in some cases and I've seen it be a complete disaster in other cases. So, you know, it sounds like you guys are still friends. That's great. Good to see that, you know. So you survived writing a book together. So how did how did it work for you? Where did you kind of find routines that made it, you know, made it work? And where like where was their tension?
Jeremy Mumford: Yeah, I think ⁓ I'll take this one again if that's okay, Jacob. I think for us it was finding time each week to visit each other's houses and just sit down and write the book. What would often happen is ⁓ Jacob would come over to my place one or two times a week and we would cast Google Docs up on the TV and we would vigorously debate a concept or an idea and Just kind of imagine the two of us just pacing around the living room, gesturing at the screen, getting really excited about ideas. And then we'd maybe pause and we'd like each silently write for thirty minutes. And then we'd reconvene and we'd kind of do it all over again. And so there's a lot of rounds of that. And then often what it is is like one person would write a chapter and then they would pass it off to the other person and that person would go through the entire chapter and kind of put their own polish and spin on it. And then we'd go back to the other person and that person would do that process again. And we did that for all 12 chapters. So very time consuming, but I feel like the end result became extremely polished as a result.
Jacob Miller: I will say it's a lot easier to go if I if I weren't co-authoring it, I think A, the quality would have been worse because like having another person to just be like, Are you sure about that? Just someone else to prod that and like that's kind of a bad idea. which Jeremy does in an excellent diplomatic way, but he's very good at communicating, like, are you sure that's what we want to communicate on one part. And then the second is that there were certain areas of the book where each of us were bringing expertise. So certain chapters, it was like, ⁓ Jeremy's leading on this in chapter nine, because he just knows it better than I do. But I'm over here in chapter five and I know chapter five and I know that content better than Jeremy does. That doesn't mean Jeremy doesn't contribute to chapter five, but there's kind of like a designated lead for that chapter because it follows their own subject matter expertise. So I I think having that as a natural lead proportion and then going back and forth was a a super healthy dynamic. I mean we're we're still friends. We still I think we're more friends now than we were before even. And so we succeeded, I think.
Jeremy Mumford: Okay.
Jess Haberman: Present. That's great. Yeah, I think ⁓ collaboration is really important even when you are the s a sole author, ⁓ to have tech reviewers who can really kind of push against some of your ⁓ assumptions and can can make your help you make your points better ⁓ is is really important. ⁓ but that's great that you guys had kind of a like whiteboarding slash writing retreat ⁓ kind of structure. That sounds a lot of fun. ⁓ I like that. I I like that a lot. and then I think we all know ⁓ publish. one book doesn't make you rich overnight. ⁓ you know, but Jeremy, you mentioned that, you know, it it was worth it, it it's still valuable. ⁓ tell, you know, what do you what are your I know it's early on in the in the process of of the you know publication, but what are your kind of expectations going in and ⁓ what are you hoping that this book will enable for you both in terms of your careers and your and and and future work and things like that.
Jeremy Mumford: Yeah. I mean it should hopefully come as no surprise to anybody that writing a book is not a get rich quick scheme. you know. But I think ⁓ we went into it, you know, eyes wide open with, you know, the purpose of us writing this book is ⁓ I think one, Jacob shared that he had some ideas that he was very passionate about, ⁓ and he wanted to share those with the world. And I think that should really be like the root reason for anybody wanting to write a book is you have something you want to share. With a part of the world. ⁓ I think two, I think this will benefit us moving forward. I think there will be benefits to our personal knowledge and understanding of AI. I feel like both Jacob and myself have really honed our understanding of how AI works and at an organizational level to such a degree, ⁓ because of writing the book. And then of course, there's all the benefits that come from being able to say that you wrote a book on a subject like this. There there are benefits that come from that. ⁓ but I think for me it's it was really the process of writing the book and the knowledge gained from it. That is probably the bigger factor for me. And then quite selfishly the sense of pride of being able to say I'm an author is very fun as well. Being able to tell people, ⁓ yeah, I wrote a book. That's pretty cool. You know.
Jacob Miller: I'll add that it does require a certain level of hubris to in art or as an author or anything like that because there are a million, you know, content creators out there. And especially if you're focusing on something deep, you you know, you're not just throwing out some 45 second TikTok. Like this is something that's going to take like hundreds of hours of your time that you're going to dedicate to this. And it takes a certain level of like hubris to say, like, well, I have something to say and the world doesn't have it otherwise. And to be able to push that forward. But that's where that passion comes from, because you know, the world I love the quote, the world is a museum of passion projects. The world is full of these passion projects that people have pursued. And I don't think it is accept you know, doesn't mean you're exceptionally full of like hubris. If you're writing something that you're passionate about, because you're bringing that passion, that depth of expertise to share with everybody. And I would add that the process does force clarity, like that discipline and clarity. And so I I love the line, I think it was Paul Graham who's popularized it, but it's to write is to think is writing forces you to clarify your thoughts and put them on paper and adjust go through it slowly and carefully. I think that's one of the drawbacks of AI writing, actually, is like, ⁓ I had some general concepts, then AI filled it out. the pressure process of like AI is tremendously hu useful. It was very useful for us to help us with the writing, but you going through the process and pain of writing and the trade-offs and what we truly mean by every word is something that sharpens your understanding of any field.
Jess Haberman: Yeah, I love that. Yeah, I I do think ⁓ publishing a book does require a little bit of hubris a lot of curiosity and patience. ⁓ so yeah, it's that that's a great lesson. you know, are there any things that for you know, for other people out there who are interested in writing a a technical or a business book, ⁓ are there any things that you would warn them, ⁓ and now that you've been through the process of of writing your own?
Jacob Miller: I would start with it's going to be take a lot of time. If you're not willing to dedicate the time and have the passion behind it. Like if you're not to Jeremy's point from earlier, if you don't have the passion behind it, you might not make it through. Because if you know, on a per hour basis of like how much money do you make out of this, it's probably not some astronomical figure compared to what you could be doing for your day job or consulting or doing something else. It's really driven by this underlying passion and drive. I would say that's one part
Jeremy Mumford: Yeah, I think heads up on the time, like buckle up, you're you're signing for a lot of hours. I think that's probably the biggest gotcha. ⁓
Jacob Miller: I I would say as first time authors having your your help, Jess, throughout the process for as a first time author, that's extremely helpful. probably like second or third or fourth, maybe like they've kind of figured it out at that point and kinda know the ropes and know their way through it. But it is pretty difficult as first time authors 'cause there's there's so much in the domain that like even if you ask AI to help you do something, it's not navigating like those relationships and the expectations and what are the you know, what are the incentives and objectives of the people on the other side of the table. It's really hard to understand what that is without some type of assistance and help.
Jess Haberman: It is interesting that publishing is centuries old and ⁓ AI still isn't trained on enough material to help you do do it step by step. It's it's such a unique process and it is so based on relationships and you know how you navigate them. So that's really interesting. so and and then I I'm curious kind of like what do you what do you guys read? What kind of ⁓ content do you take in as your ⁓ you know, what informed your your writing and ⁓ did you were you inspired by other books or other pieces of media ⁓ that sort of inspired the work that you did?
Jeremy Mumford: Yeah, I think as far as reading, ⁓ for me there's kind of two sides to my reading life. There's like books and content I read for fun. ⁓ that's a lot of like Brandon Sanderson and like Golden Age Science Fiction. ⁓ a whole range of books, you know, I I absolutely love reading. ⁓ then on on the more technical side, ⁓ I do like to get into technical books and we had several that we we looked to for examples. ⁓ You know, there's books like Observability that wa that was really fantastic, ⁓ and a few others. LM can it's the LM Engineers Handbook was another favorite of mine. ⁓ I think that one's by Paul Utsin. ⁓ that was kind of a helpful example of writing for AI in the modern age. ⁓ and we would reach out to these authors on on LinkedIn and other platforms and ask them for advice and say, What would you advise about writing? And so, you know, don't be afraid to Hit people up. it's a very friendly community. People love to share ideas. You know, we even got some local Utah authors that we talked to and asked for advice on which publishers they preferred and ⁓ what writing techniques and styles they recommend and habits they would recommend for us. So all of that was very helpful.
Jacob Miller: Yeah, Jeremy highlighted, for example, observability was one book. There were certain books that we saw within the text space that were at least a few years old that were still hyper relevant over time. And especially certain chapters of them that were hyper relevant. And so like you'd you'd read this chapter, you know, within a larger book, and like you could have read it fifty years ago or ten years ago or two years ago or today or five years from now. And it's like it still resonates, it's still true. And that was like That's w exactly what we were shooting for, but with the audience that we had in mind and then on the topic that we cared about. So combining those three, there were these lots of like these specific examples, and they usually were like a few chapters within a larger book where we'd you know, we'd read those books and they really stuck with us over many, many years.
Jess Haberman: So ⁓ so what's next for you? Are you thinking about writing another book after this? So you need a break or ⁓ are you ready ready to keep keep going?
Jeremy Mumford: I mean, I need a break. It's that was a lot of work. I think next steps for us, obviously, right now at the time of recording, the book is still under pre order by the time this comes out. It'll be available on Amazon and Barnes and Nobles. And I think we're excited to go out and share it with the world after having worked on it so hard over the last year. We're really excited for people to dive into it, see what they think. ⁓ And then we're busy kind of applying all of those concepts we talked about at pattern. So, you know, we are practitioners of what we do right now. We're we're very busy building a lot of stuff. ⁓ and that's a lot of fun too. So that's kind of the immediate future for for both of us.
Jacob Miller: Yeah, we've got a lot of building, some consulting and some helping as part of that, but also some conferences, stuff like podcasts to share those the key learnings and key concepts and ideas. ⁓ there's there has been a book that I've been like working on kind of within the AI space, but as a little bit more general audience as far as the lessons that you need to know in interacting with AI systems as like regular people, with the more general audience, but that's someday maybe that will see the light of day. For now it's it's essays that I post on on the blog and things like that.
Jess Haberman: Yeah. That's brilliant. ⁓ I mean it it it is impressive that you published this whole book and had a full-time job in the meantime. ⁓ so congrats on that. Major kudos. ⁓ yeah, and I think like I tell authors all the time that get you know, getting the book published is a huge feat and it's worth celebrating, but it's just step one. It's there's you know, you gotta get the word out. You've invested so much time and energy in writing and then, you know, it's it does require still a lot more energy to get it in front of people. So ⁓ so don't be shy. definitely make sure everybody knows about it. Hit them over the head with it. ⁓ because it's really important. It's so timely too. I think it's a great topic that everyone's already talking about. ⁓ So yeah. Is there anything else? I you Jacob, you mentioned your ⁓ your blog, but is there anything else you want to plug? Where can folks find out more about you?
Jacob Miller: Yeah, so we write at the whybehindai dot com. It's where we post essays, we keep things a little bit lighter there. Many concepts from the from our book end up there or have ended up there, although the blog form is a little bit lighter and more entertaining and has better memes, to be honest. And then the book itself is Architected Intelligence, and you can find us on LinkedIn, Jacob Miller and Jeremy Mumford. Did I miss anything, Jeremy?
Jess Haberman: Amazing.
Jeremy Mumford: I think that's it.
Jess Haberman: Awesome. Well, thank you both for joining the pod today. It's been a pleasure speaking to you and hearing more about the book and and the whole process you went through. Congrats again. ⁓ please everyone do order a copy of Architected Intelligence available where books are sold. And ⁓ hope to talk to you guys again soon. I can't wait to see where this goes, and I hope all the success for your book.
Jacob Miller: Thanks, Jess.
Jeremy Mumford: Thank you.
