Hey everyone and welcome to today's episode of Developer Team. My name is Jonathan Patrell. My goal on the show is to help driven developers like you find clarity, perspective, and purpose in their careers. But today is a little different. We're going to do something a little bit different in this episode. And we're going to talk a little bit about some of the stuff that I've been using, some of the things I've been seeing and enjoying recently. And I want to do this for a couple of reasons. One, the industry is a little bit crazy right now. And you're probably overloaded with every other episode, every other podcast talking about how behind you are. And trying to pick the things that you're going to invest your time in can be really hard. and sometimes it helps to see what some other person is doing. Sometimes it helps to just get a glimpse into one other person's viewpoint rather than looking at the 30 new tools that you need to be adopting and getting overwhelmed with all of the blog posts that you just don't have time to read. Sometimes it can be grounding to just think about what what one other person is doing. And so that's my hope, part of my hope in today's episode. Another part of this is to kind of give you an idea about how someone in the industry like me, how I'm thinking about some of the problems that we're facing as engineers, but also Also, I think that so much of the benefit of human-created content, which all of the content that I produce on this show is stuff that I have written, or at least the vast majority is coming directly from my brain. And certainly when I'm speaking to a camera like this, I've written all of this. I think it's helpful to see what another person is actually engaging with rather than only engaging with algorithmic kind of production. And I want to be clear, like nothing that I'm talking about today, except for our sponsor today, nothing else is sponsored. I don't have affiliate links or anything like that. This is just kind of how my workflow and beyond my workflow, my personal life, how I'm engaging with different tools and services and things that I'm interested in. And hopefully this is a useful kind of picture into the world of an engineering manager in the spring of 2026. It's very possible that you're listening to this much later and you're going to look back and say, wow, you're way out of date. uh hopefully hopefully those of you who are listening when this releases uh it doesn't feel that way so uh you know part of my part of what makes me feel fulfilled and alive uh are are my various interests and uh you know one of those interests is is flying and one of the things we use as pilots is checklists. And I've decided recently to try and experiment where I'm actually using physical checklists for my own life beyond the cockpit. So I've made these checklists and I've printed them out. I even have them in like a physical, like an actual pilot checklist, list, you know, kind of flip book. And it's like an eight and a half by five and a half, you know, sheet of paper. And the goal is to have, you know, the things that I do repetitively, including, you know, producing this show, the things that I do predictably to have checklists for those things. And I have them for some of my work events, things like one-on-ones, things like stand up, have these checklists. And some of this is just my own kind of personal experimentation. What is it going to feel like to have a checklist? You know, I've always enjoyed the idea of desk references. I'm not sure if this is something that is unique to me or to a subset of people or if this is a universal thing. But I think that checklists, at least for me, provide provide me a good touch point, a good reminder to, because my working memory may not necessarily hold on to, okay, you're supposed to do all seven of these things during a, you know, a standup or something. And so for, if you're like me and you have a lot of different things that you do on a repetitive, like a repeated basis, then having checklists might be a useful thing. This is inspired by a handful of things. One, as I already mentioned, pilots use these all the time. There's the book by Atul Gawande, I'm hoping I'm saying that correctly, called The Checklist Manifesto, where he talks about these being used by physicians as well. They adopted the practice from pilots, partially because aviation has such a strong positive safety record. record. It's also inspired by a book called Noise by Danny Kahneman. Danny Kahneman is the man who wrote Thinking Fast and Slow, but he also wrote his second kind of major book was Noise. And Noise talks a good bit about how procedural kind of algorithmic approaches to things can reduce variance, unnecessary variance in the outcomes. And he talks about that in Thinking Fast and Slow as well. To cap all this off, I've been thinking a lot about how my communications and my thinking, and we'll talk more about AI stuff later in the episode, but my thinking as I'm building skills skills for Claude or as I'm providing the context to an LLM, that a lot of those skills are transferable into my communications with other people. That if I had more precise communication skills, if I was more precise in my way of thinking about operations, then perhaps there is going to be a positive carryover in the way I approach my work. And so this is part of the the reason why I decided to kind of experiment with this checklist approach is because it does increase the kind of rigor or, you know, the amount of energy put into being clear and following this more procedural way of doing, you know, repetitive things. Another simple example of this is cleaning the turtle tank that we have. We have a turtle and a big tank and a big filter for the tank and, And every X number of days or months, we need to clean it in a particular way. And those checklists are going to be a little bit different depending on where in the cycle you are. All of these things are as simple as individual steps. As kind of a bonus, I actually created a quad skill. Again, we'll talk about AI later in the episode, but just to give you an idea of how I'm thinking about how I incorporate AI into my work and into my personal life. I trained a cloud skill to be able to create these checklists in this specific format. So now I can just ask it to do that and it will provide that in like a print ready form. So that's one thing I've been experimenting with. I've also been experimenting with, let's see, kind of information architecture changes in my life. So really what this comes down to is changing our physical spaces. We moved into our current house about coming up on six years ago, and we built this house. And, you know, over time, my wife and I, we've acquired, you know, our variety of things that we've acquired. I am a serial collector of a lot of different hobbies. I have a lot of things that I've collected as a result of those hobbies. And parts and pieces of those things have kind of been scattered throughout the house. We've not had a great organizational structure. And as I've been thinking about increasing the rigor in my communications and in the other parts of my life, the physical checklist, et cetera, I realized that this organization And putting effort into organization is something that pays you back many times over and over. And the metaphor that I'm using for this is that reorganizing is essentially like creating a very fast index, right? You know where things are and so you can find them more easily. You know where they go so that when you are cleaning your house or you're straightening up or whatever, you can return them to their location. You know the things that you have available. You know what you own because everything has a location to be. As a part of this, you know, one of the simple things that I've been using is a brother brand P-Touch label maker. Highly recommend this. A very simple adjustment in my life is to put more labels on things. And you'd think, OK, this is not a revolutionary change. Right. And that's the whole point of this particular episode is to not talk about revolutionary changes. the small things actually make a big difference in our lives and having labels on things has actually improved my life and in a meaningful way. So, you know, I guess the point isn't for you to go out and buy a brother P touch label maker and do exactly what I did, but instead to think about what are the contours and the small interventions that you could have that might have that same kind of improvement for you. If you're looking around you and you find that, you know, having more rigor in your approach to something may help you, then how does that carry over into other parts of how you're operating your life? Another cool thing, you know, I've been into 3D printing for many years, but my wife for Christmas bought me the Bamboo P1S. And this particular printer is a significant upgrade to my previous printer, which is a Prusa MK3 or MK3S. And the bamboo is much faster. It's simpler to operate. There's kind of a whole kind of community of people who have published files for printing. And this has also helped with some of the organization efforts. So everything kind of has worked together to improve organization. And a totally different sphere. here, part of what I've been trying to do in my personal life, what I've been trying to improve at in my personal life, is staying on task. I actually literally have a book that I'm reading right now called On Task that talks about executive function and some of the more, currently it's talking about the anatomical, the brain anatomy and where does executive executive function actually live. But as a part of this, I've stumbled onto a very useful intervention that I didn't realize would be as effective as it was. I have a home gym. I always thought that having a home gym would make my workouts very efficient, that I could go into the gym, spend the time that I would need to spend, and then I'm done, and there's no travel required. You know, I can, I can break up my workout into multiple sessions if I need to, et cetera. Right. And a few months ago, my brother-in-law asked if I wanted to join him at a local gym that's down the road, just Planet Fitness membership. And at first I, you know, having done powerlifting for a few years, I kind of wasn't interested in the idea of going to this type of gym, but I knew I was also going into, and this is kind of like insider baseball discussion, but I was going to go into a body recomposition effort where I'm going to lose some fat and improve my muscle mass and et cetera. So I knew that I wasn't going to be able to continue my normal workout schedule. And I decided, okay, I'll try this out. I like my brother-in-law. He's a good guy. I'll enjoy working out with him. Maybe I'll learn something about a different style of workout. This has been a huge improvement in my workout focus. A lot of times when I would work out alone at home, I would end up getting distracted. I would get off task. I would procrastinate. I would end up on my phone or even bringing my computer into the gym and doing work instead of actually working out. And so all the kind of intermingling of that context has made it more difficult for me to kind of build the habit of sticking to the workout. And so going down the road and working out with my brother-in-law has been one of the best productivity and focus interventions for my workouts. workouts, you know, if, if you're kind of struggling with this kind of thing, like I am, then I highly recommend James Clear's book, Atomic Habits. I'm sure we've talked about this on the show before. Atomic Habits is, is kind of like the, I don't know, my, my top recommendation for habit building in terms of, you know, the science and how to actually do it from a tactical standpoint, how are you going to build a new habit? So in the meantime, I'm going to work on developing the right triggers using the information from Atomic Habits to help me, you know, once I move away from working out in that external environment and back into my home gym, that I won't, you know, lose some of those improvements. improvements. Speaking of books, a handful of books that I've been reading that I think have been useful to me that are not necessarily on the same train, but have kind of triggered some interesting thought, at least for me, about my job and about the current environment. One is Richard Rumelt's Good Strategy, Bad Strategy. I'm following up by reading The Crux. most of my reading is done on Audible by the way so I'm just listening to these books while I'm doing other things you know if you really need to soak in a book I recommend doing both actually for me listening and reading you know reading either an e-book or a physical book is the best way for me to actually get kind of both or the best experience I know that it could be an expensive option so I start with with listening and if If something really sticks with me, I may go get the e-book option. So Richard Rimmel's Good Strategy, Bad Strategy. This is an excellent book to even think about, kind of define what strategy is. Probably the most important insight for me has been that a good strategy is focused on a problem. and a solution that is unique. So anyway, I've enjoyed Good Strategy, Bad Strategy. The Crux is supposedly kind of a more focused version or it covers some of the same kinds of information, but from a more focused angle. I also very much enjoyed Bungay's Art of Action. This is a great book. It is steeped in a lot of kind of military language. So if that's not your thing, then you may want to skip it. But I think this particular book could be very useful as a way of thinking about dealing with agentic workflows. And again, we'll talk about AI in the latter half of the episode. So, but Bungay talks about how to translate intent from multiple levels up all the way down to, you know, the front lines essentially, right? And it is, you know, it's based or I guess it's adopted from a set of military strategies that ultimately did very good, did a great job at adjusting to new information and new restrictions, new restraints in the environment. and bringing the intent from the top level all the way to the bottom level of an organization. I'm also enjoying, on a totally different kind of flipping the page here a little bit, I'm enjoying the book How to Know a Person. This is by David Brooks. And it really talks about how to avoid things like typecasting people And instead to see people in the unique and true light that they are. You know, this is something I'm always at a personal level, my personal values. I'm always investing in my personal relationships. And this is something I care very deeply about. So if you are like me and you are interested in, you know, developing really deep relationships, not necessarily a million of them, but a smaller number of deep relationships is what I tend to do, then how to know a person will kind of guide you on that path a little bit. As I approach, I'm approaching the age of 40, and there's this book called 4,000 Weeks, which is subtitled Time Management for Mere Mortals. This is probably my number one recommendation. If you read nothing else on my list here, this is such an excellent book and it will help you realize that the overwhelm that you may be feeling is, first of all, very human, very normal to feel, but also that no one has a magical solution for you. There is no special thing you can do or special productivity hack that you could employ to get ahead of all of this. And so if you are looking for a book that kind of reminds you to think about what's important in your life, then 4,000 Weeks is an excellent book to pick up. This is one that I actually did read the ebook version of. Highly recommend it. But something that I like to do on a regular basis is to learn something that I'm fundamentally new at, or at least close to new at. Right now, that thing is learning to drive a stick shift again. So as a Christmas gift to my wife, I found an old used beat up Bronco. And I say it's beat up. It actually isn't. The body's in pretty good condition, but it is old. 1983 model. And so my wife and I are spending, you know, the evenings, you know, kind of taking it out on a back road and learning how to drive a stick shift. And this is, you know, like most new things that I've done, reminds me of the value of intentional practice, of kind of going through a very specific set of actions and improving on those actions, and then going back through and refining and improving and refining and improving, setting goals for how many stalls, for example, are you shooting for in this particular ride? Driving stick shift is not an extremely complex thing, but it is something new and becoming good at driving a stick shift. This is reinforcing this habit, learning as a habit, not necessarily that I care so much much about driving a stick shift, but rather that I care about continuously being in the beginner's seat and learning something new all the time. So those are some of the things that I'm kind of engaging with. I want to kind of go on a quick list. This is very unusual for the show. And frankly, it's almost out of character for the show. But I think it's an interesting thing to talk about. I'm going to talk about some of the things that we've been watching and really enjoying in our recent time that I think are really engaging media for us. So this is kind of just a fun list. We've really enjoyed Night of the Seven Kingdoms. I think the character development that Night of the Seven Kingdoms does is really excellent. If you watch the original Game of Thrones. This has a totally different tone. And, you know, maybe mild spoiler alerts here if you're planning on watching it. But Night of the Seven Kingdoms, the really interesting thing that it does is it kind of subverts your expectations. It changes the format completely, you know, or while staying in the same kind of cultural setting, it changes the format and the the tone of the show, and it feels like kind of a different thing. So I think they did a really good job with that. We're going to embark on re-watching Silicon Valley. And some of the reason for this is because it seems to have been nearly prophetic to her current moment in time. And I just think it was such an excellent series. Start to finish, we really loved Silicon Valley, our first watch through. It will feel even more interesting and maybe even nostalgic at this point to rewatch Silicon Valley. That's on HBO. Another show that we enjoy just as kind of our zone out show, we have a couple of those. One is Taskmaster on YouTube. This is an excellent show. I think engineers would love this show, generally speaking, because it is full of various puzzles and things that could be engineered throughout the show. It is very entertaining. I highly recommend it. A couple other shows that we've enjoyed recently. One is Shrinking on Apple TV. There's a second season out. we're big fans of Shrinking the first season and looking forward to the second season. I guess we're on the third season now. But anyway, Shrinking has been excellent all the way through. I think it's a very good kind of human recentering kind of show. And then we've also been watching a lot of Dropout. So if you're not subscribed to Dropout.tv our favorite Favorite shows right now are Make Some Noise and Very Important People. Those are ongoing. But then there's also Game Changer is another very popular one from Dropout. You may have seen clips of that. But we've been enjoying those a lot. So you might be wondering, OK, Jonathan, you're talking about TV shows. You're talking about books. You're talking about all these things. But you're not talking about AI. That's what I'm here to learn about. I'm a software engineer. Why aren't we talking about AI? AI. No, I'm not abstaining from using AI. I already mentioned that we're going to talk about this in the latter half, but it's such a loaded topic. It is so overloaded that I wanted to give it its own kind of dedicated section. So we're going to talk about AI right after we talk about our sponsor today. night. Today's episode is sponsored by SERP API. If you're building an application that needs real time search data, whether that's an AI agent, an SEO tool, a price tracker, or anything else that needs to know what's happening on the web right now, SERP API is the web search API that handles all of that for you. You can make an API call. You get back clean JSON. That's it. They deal with all the proxies, the captchas, the parsing, all the scraping headaches that you don't ever want to think about. And if you've been doing scraping recently, then I feel for you. They support dozens of search engines and platforms. They're fast and they're good enough for NVIDIA, Adobe, and Shopify to all rely on them. You probably can too. There's a free tier to get started so you You can build your test application before you commit to anything. And by the way, if you are building with AI, SERP API has an official MCP to make getting up and running a simple task. If your app needs to search the web in real time, check out SERPAPI.com. That's S-E-R-P-A-P-I dot com. Thank you again to SERP API for sponsoring today's episode. So AI is such a big topic. I want to kind of cover some of the ways I'm thinking about AI and some specific tooling. But really this portion of the episode, I want to focus on how to think about this, the current environment and how AI sits against all of our jobs and how we can respond to that. that. So first, I want to say that we're definitely well beyond cat out of the bag scenario here. If you are working in software engineering, you almost without a doubt will be using AI tooling, agentic tooling, you know, something that is generating code, something that is tying all all of these services together, that's very likely a core part of your skills and tools going forward. That can be an uncomfortable thing because it means a lot of change. It means a lot of change. And we'll talk a little bit more about this later, but change has its own skills. skills. Being flexible and learning are their own skills. Being able to move from one environment to a new one, adaptability, being able to change your mind, being able to update your thinking, right? A very simple way of thinking or example of this is if you tried agentic workflows, flows, if you tried cloud code in October of last year, and then you try it now, the experience is drastically different because the underlying technology has changed significantly. And this stuff is changing very quickly. And so being able to deal with change, these are all core skills there. And this is new. This is new for us that our core skills need to shift this dramatically, this quickly. You know, if you look back over your career, there have been some very large changes like this. If you think back, you know, moving from a physical data center to the cloud, this was probably a massive mindset shift for a lot of people. You know, going from slow internet to fast internet, going from no internet to having internet in the first place. These are all seismic shifts that totally changed the landscape for software engineers, totally changed it. Going from having only compiled code to having dynamically run code, pretty big shift. It may not be as broad impact, but it certainly required people to learn something new. And so we're in a situation where change is happening, but it's happening at a rapid pace because the change itself is feeding back into the rate of change. The change itself meaning that we're using these new innovative tools in ways that humans on their own wouldn't have the ability to move that quickly. but because we're less reliant on humans in order to make this change accelerate, the change is going to accelerate faster than what we're kind of naturally used to seeing. So this is the core skill of being able to adapt to change. That is becoming more and more important. Okay. So we'll talk a little bit more about how you can think about where this sits. I want to talk about some of the things I'm using right now. I use Cloud Code. I actually started using Cloud Code, I think, the day it was released, maybe the next day or certainly that week. And it has significantly improved. Those early days still kind of felt like you were trying to remind it how to stay, how to retain context. But I wanted something that could retain context. I'd used kind of in-browser, for example, Anthropix projects and their artifacts. And I tried to, you know, and succeeded in kind of, you know, building things by using that kind of workflow. But it was never ergonomically useful. It didn't feel as good. some of the things that you know and we dealt with back then back then being just last year was the LLM kind of trying to tell you to leave the rest of the code the same and then swap out this one part right so that's kind of that's where we were less than a year ago so now you know I've I've started using cloud code. I still use LLM chat interfaces for simpler questions, things where I'm away from my computer, use a one-off, maybe image generation or doing something with an image that I've taken. But this happens less and less each day. For me, the most powerful use cases for cloud code, one-shotting things, right? One-shotting tools that have very clear boundaries and expectations for the implementation. Those things can be created in one or two prompts and most of what they need to do is complete. Once that promise is given and once it's iterated a little bit. So doing those one-shot things. Another really useful thing that I found Cloud Code for, using Cloud Code for is, you know i used to think maybe i'll do that one day or maybe i'll make that improvement i'll make that environment improvement you know my my local environment oh i'd like to have that on my my terminal prompt for example well now i can just do that right i can do it almost as quickly as i can think about it almost as quickly as i could write it down as a note to remind myself later now i can essentially feed that to cloud code and almost immediately i can have something that is close to operational, close to working, you know, nearly immediately. So these are like environment improvements, nice things that are, that I know are technically possible, right? Things that I know are, you know, if I had the time, I could do them, right? That's the kind of stuff that is very easy for me to hand to Claude. And, and it does a really good job of accomplishing those. And by the way, this is not, you know, again, none of these things are sponsors. I don't know how the kind of IDE type agentic flows, how well they work. I imagine they have similar levels of quality. Codex, for example. So I've just chosen Cloud Code because I've really enjoyed Anthropix, the quality of Anthropic since I started using it. So I've also been using Cloud Code as kind of a terminal operations manager. So if I have a problem, problem, you know, with a package or something, I ask Claude to figure out what's going on. Sometimes I'll ask Claude to install something like find the latest version and install it or find the version that's compatible with my OS and install it. And if you run into any issues, then work through it. And a lot of times, you know, things that like maybe I was missing a header file or something, it will figure that out and fix it for me. You know, I don't necessarily recommend all of these ways of operating as some of the things that I do with cloud code, maybe a better engineer may not do, right? And I think that's part of the spirit of this episode is actually to remind you that there's not really one right way to do these things. It's changing so quickly and everybody's going to have different recommendations and reasons why, you know, people listen to this and say that it was like some of the things I'm doing are terrible ideas. And that's totally okay. I actually encourage you to kind of develop those thoughts and figure out what's right for you. I'm using OpenClaw, which I'll talk about a little bit more in a minute. So in order to access OpenClaw from all my devices in a secure way, I'm using Tailscale, which is free for me, at least it's, it's been free. I don't know if it's still free or if it will continue to be free, but it was free at the time that I recorded this to access my, you know, essentially all of my devices. I'll give an internal IP essentially like a tunnel, I guess. I'm not entirely sure how, how it's actually working. We're using a pattern called spec spec-driven development at work at Gartner. And this has shown a lot of promise. I recommend Googling this. I'll probably include it in the notes as well. If you are using something like Cloud Code or some other agentic workflow for developing, for generating your code, then spec-driven development is showing a lot of promise. It is essentially providing the right context, shaping context, shaping the input, shaping requirements. in order to produce the right kind of output, right? So that is really the goal of Spectrum development. And there's also some like test loops that you can implement as a part of the Spectrum development, ensuring that certain requirements have been met. So that's some of the things that I've been doing with Cloud Code. I've also fixed a handful of developer T issues. I've been doing some back office support work, like finally getting transcriptions, built for all episodes. So we should have every episode on developer t.com should have a transcription, it's not gonna be perfect, because there's over 1200 of them. And we're kind of automatically transcribing these episodes using whisper. But there are it should make the show much more accessible, it should make it easier to eventually to build a more powerful search. and for me to kind of dive back into content that I've forgotten that I even created you know 10 years ago is kind of the goal of that so I mentioned I've been using open claw open claw is very much experimental it's not something I'm relying on for much of anything important at all but for some of the fun things and like experimenting getting on the edge of you know know, what's possible with some of these workflows. You use OpenClaw primarily through a chat interface and OpenClaw will then go and develop things on your behalf. One of them that I've built is an agent for trip planning. And what's cool about it is, you know, you're not building skills for others to consume. This is something that you can make really specific to you. And so my wife, my kids and I, we tend to take, you know, only a handful of types of Some of them are family trips. Some of them are trips to the beach. Some of them are trips to like a theme park. And then some of them are, you know, trips that are more like day trips, right? We're going to go stay, you know, one night somewhere and then come back. And each of those has their own concerns and considerations. You know, my wife, for example, has a gluten allergy. I like to find a good gym to work out when I'm, you know, on a trip like that. And so those specific things I end up having usually to go and do my own research on. So a typical trip planning software workflow is not going to have all of that stuff built in. I can just make the agent do all of those things that are really specific to us and build our packing list and all of these things all through a simple chat interface. The agent is specifically dedicated to that context, and it does a really good job at finding things that actually will work well for us. So we've tried it on one trip. It worked pretty well. We're going to keep refining that agent and seeing how well it goes. I already mentioned the podcast transcription automation. I've got some things that OpenClaw is doing to enable that. One thing that's really cool. This is just a fun thing, kind of a nerd out for a moment. It's Home Assistant. I have Home Assistant for a lot of the things that we run around the house. For example, our garage door. We have like a simple sensor that tells us if our garage door is open. And then we also added like a read switch and an automation so that I can open and close it remotely. Well, these automations can be a little bit fiddly to create, but there is a very clear kind of YAML format for them. So this is actually really good for an LLM. So I could talk through my OpenClaw to create new automations and it directly talks to my home assistant box and it knows how to do new automations. And that is working incredibly well. So if you have a home assistant instance and you enjoy the automations, but you'd like like to create some more, then you may look into using an LLM for that or using OpenClaw more specifically so that it can talk directly over your local API. I've also been using Cloud Cowork, which is kind of the non-code version of the Cloud kind of agent runner local system thing. And Cloud's done a very good job or Anthropics done a good job of making these things have of productized. The cloud co-work does a pretty good job, you know, handling things like file system improvements, like cleaning up your, your desktop, you know, for you. And, and, you know, one, I think one of the built-in automations is to look at all of your screenshots one at a time and evaluate what's in them and then put them into folders based on the contents of whatever's of is in the screenshot. So cowork is pretty cool for that kind of thing. I've been using cowork and quad code and open call in combination with obsidian. Obsidian is just markdown files, but it's kind of a layer on top of your markdown files. And I use this for knowledge base. All of my meeting notes go in there, you know, list to do lists end up in there. And, you know, task management, I handle in Todoist, but Obsidian, I tend to kind of use as my notepad for capturing things that are kind of in process. If I have something to do really quick, like three or four things I'm trying to get done in the next couple of hours, I might throw them in Obsidian so I don't forget them. It works really well as like a context hybrid as well. So I can allow allow cloud code access to read context out of my obsidian to then, uh, you know, use that context to go and do work somewhere else. Right. Um, I can, I've, I've also been using obsidian to catalog things about my personal life and surroundings. And my belief here is that the more kind of padding and context and personal information, the kind of personal information that I want my LLMs and agents to have about me, the more that it has about me, the more it can answer and do on my behalf, and the more kind of synthesizing it can do on my behalf as well. Well, as a totally different note, or on a totally different note, there are a lot of powerful things that I've been exploring that I can use these tools outside of my work and software skills. For example, I connected an MCP for Fusion 360, which is a 3D modeling software. I already mentioned that I'm into 3D printing. And this actually does a pretty good job. I found, I can't remember who published it, but there's a cloud skill that you can install and use that cloud skill in concert with the MCP to design parts like functional parts that are parametric design, which means essentially you're designing using measurements as your kind of core. and this has been a really interesting process because i'm able to design things at least the first version of it entirely by providing the specifications um in in plain language to in this case uh clogged co-work and co-work knows how to use the mcp and uh it also knows how to like take take screenshots inside a Fusion 360 and check its work. And it does a fairly good job of producing, you know, the kinds of models that you would want. It's something that will get better over time, I'm sure. And sometimes it has to iterate multiple times. It has a habit of starting fresh and then rebuilding from, you know, the very beginning rather than going back into the timeline and figuring out where, you know, to make an adjustment. But overall, I'm very encouraged by this because I don't have a lot of time myself to go and learn a lot of the more meticulous skills of a totally new habit or set of tooling. tooling. And while I could make this one of my new things I'm learning, I have learned enough about Fusion 360 and parametric modeling to do most of the things that I care about. But sometimes I run into that same situation where it's like, well, I could probably figure it out, but it's not really worth that much of my time. And I could just go buy a commercial product to fulfill that need. Now, AI has kind of brought that a little bit closer, right? There's a little bit of a bigger circle. And this is kind of a principle that extends beyond 3D printing, certainly. There's a little bit of a bigger circle of things that I can accomplish on my own, or things that I can accomplish that previously I would have just ignored or not done at all. So I actually think that's a really empowering way of thinking about AI. And I want to think for a moment out loud about how I'm thinking about AI and its use cases. Certainly, that is one of the most important things to consider. And that is that the agency that any one person has, the domains that any one person can work in, the kinds of things that any one person can do, that has extended a little bit, right? We have a much different picture of what we can accomplish with the tools at our disposal. And so, you know, as much as creating simpler to read programming languages extended extended the capabilities of the average programmer, AI is continuing to do that, right? We have a little bit more capacity. And now the fundamental skill is being able to describe, right? Using language and being clear and understanding how these models works that we can use language in a way that is appropriate for the models, right? So I'm thinking about AI AI and energetic flows in the following couple of ways, different categories here. The first category is automation and improvement of existing workflows. So this essentially is removing friction to action. Most of our AI efficiency gains are happening in this category right now, right? We are essentially kind of overlaying AI to make the existing stuff that we do easier, faster. And because of this, we're kind of freeing up resources. We're finding new bottlenecks. People are moving faster, trying to figure out how do we make sure that our teams have enough work on their plates. It seems like a really exciting thing. but mostly it is just taking the existing stuff and making it more efficient. That is the biggest change a lot of people are seeing right now. But that's quickly rolling into improving our actual ops, right? Operation, polish, and kind of adjusting the way we think about operations. What are we doing that we don't need to be doing at all, right? What kinds of functions do we have on our teams that we could actually replace with totally different functions and get a lot more done or create more value in the long run. And so you could think about this as, why do we have support systems for things that we no longer need support systems for? That's one category. Another category is, how do we make sure that we are responding to those new emerging bottlenecks doing something about it. Previously, when we identified a bottleneck, the ability to address that bottleneck was very slow. It's a human process. A lot of times our teams would adapt to the bottleneck size, for example, right? But now if we can use AI to help us identify the bottleneck, then we can also use AI to help us relieve the bottleneck in many cases. So that's likely what's going to happen is we're going to see operational adjustments. This really comes down to what happens when we remove humans from this particular functional area. And to be clear, what I'm not saying is taking humans out of the equation entirely. What I am saying is freeing humans up from having to do that particular task and allowing them to do something different. different right they can go and do a different task well now if we have a lot of those more expensive operational things removed humans removed from those what is the highest value next thing that they could go do right because um the it's very important to recognize that if we remove humans from from this thing and company a just takes the humans away and tries to be more efficient, but company B says, Hey, you know what? We're actually going to put those people, we're going to put those humans onto higher value things. Then the net output of company B is likely going to be better than the net output of company A. And now we have market effects. That's making company B more competitive than company A, right? So it's unlikely that, you know, in the long run, the companies that are just looking to cut costs by doing big layoffs, AI driven layoffs and that's it that's the end of the story for them well they're not going to actually compete against the companies who are redeploying those human resources into higher value places humans are still better than AI at a lot of things so until that changes the companies that are using both are probably the ones that are going to get ahead Okay, other categories, execution of complex technical tasks. This is where the agent is kind of doing the technical implementation. They are on the ground kind of, this is the front lines, let's say. A lot of code generation would happen in this category. And those kinds of actually building things, right? Or actually moving things around. out. That's the execution of complex technical tasks. Iterating on intent, iterating on outcomes, iterating on goals. This is more like a tactical consultant, right? This is a consultant that is looking at what you're trying to do and kind of adjusts the output with you, right? They're going to work with you to figure out, is that actually your goal or are you trying to do something else. And then we have kind of these more abstract things. And we haven't really seen a lot in this direction yet. But I see this certainly is on the way. And that's the exploratory connections, abstract modeling, context modeling, and creative synthesis. What this really looks looks like is drawing conclusions from things that are genuinely illuminated that you didn't see before. Now, you may say, well, we already see that. We see that Claude can look at data and draw out conclusions. Of course, that's possible today, but drawing out meaningful connections, drawing out meaningful adjustments in strategy, trying to forecast the proper kinds of decisions or direction or helping a person actually build a strategy. A strategic consultant is not something that we've seen a lot of yet, but it is something that is likely going to be part of what AI can help with in the future. AI also works very well as an encyclopedic knowledge, like a powerful kind of search agent. We've talked about that with SERP API, the sponsor today, Also know that this is something that agents can go and do on our behalf as well in our agentic workflows. Those agents can go and gather information that we have not actually gotten ourselves first. They can go and get it for us. We have execution of clear workflows, which kind of fits into the earliest bullet point. On this list, which is the idea of improved efficiency on known pathways, but if we have these clear workflows for AI to execute against, then following those instructions, this is going to be something that AI becomes very fast at. So the lower level models, what are today are frontier models, will end up becoming faster and faster. and the fast models, you know, they'll become the fast models. And so now we're looking at, you know, some pretty complex things that can happen, but the clear workflows and improving the workflow clarity is going to make those highly efficient to execute on. So what should we do with all this? What am I doing? You know, to be honest with you, I don't think anyone can answer this question completely yet. People are trying things. Some things are working really well. That doesn't mean that you should go do them. It also doesn't mean that they'll continue working well, right? So what should you do with it? You should try things, right? My main recommendation for you coming out of this, and then I'm going to give you some more specific information, but my main recommendation is to be action-oriented throughout all of this. As we said earlier in the episode, the cat's out of the bag. If you want to stay in this industry, if you want to continue working in software, if you want to continue working in innovative startup companies, there is, I would say, a nearly zero chance that you're going to be able to do that without incorporating some of these skills into your skill set. that learning about these models, learning about how they work, toying with, you know, agentic workflows and figuring out how that stuff works. It's all going to be very important. Okay. How should you spend your time, right? Things that I believe make your time using AI abundantly more useful. I'm going to give you a handful of these here at the end of this episode. The first is spending time planning. figuring out the guardrails, figuring out the ideal outcomes, figuring out the expected decision points, the freedoms, how should you expect this AI to act, where are the rules, are there any formatting templates, is there anything that this thing should do that you can specify? Spending time in that realm before you send the agents off to do their work, work that is time well spent it's abundantly more is exponentially better outcomes if you can wrangle that skill set in by the way we mentioned the book the art of action this actually is a very important part of the art of action which is essentially it's what plan mode does if you use use cloud code. Other agent tools do this as well. You ask for an intent of some kind. Your agent goes and makes a plan. It brings you the plan back. You refine the plan, right? In the art of action, this is called briefing and back briefing. So this is a pattern that works well. Well, iterating first on a plan, getting good at iterating on a plan is one of the most useful, most powerful things you can do with AI. Okay. Investing in mindful change. This is kind of flipping and looking more inward, right? What is mindful change? It really means understanding yourself. Start with understanding your values. understand how you want to respond to this moment. In 10 years, when you look back, how do you want to have acted? What do you hope you say about yourself? What do you hope others say about you? The people that you care about, the people that you love, what do you hope they say about you? Who do you want to be in a moment like this? What is your deepest identity desire? How do you want to respond to this? That is the first thing that you need to clarify in order to decide what to do next, right? And this takes, introspection is not obvious. It's not easy. It's something that you should spend time journaling or spend time, you know, talking with if you are a verbal processor or if you have really trusted people in your life, talk to them about it. But this is a very important thing to do because in times of change, not knowing who you are, not knowing what you care about, not knowing your own values, the change is extremely difficult to navigate, right? Because you don't have a GPS, a homing beacon. You don't have a rudder. You don't have an anchor. There's something that is missing because you don't know where you are today. day. And so when somebody is asking you to go somewhere new, it's hard to know if you're willing to do that and how you want to do it. Right? So this feels very abstract because it is. There's a lot of things about, you know, digging into your own values. Where do those values even come from? It's hard to articulate, but it becomes concrete. It becomes concrete in your decisions about where you work becomes concrete. Your decisions about what you're willing to do versus what you're not willing to do, right? So dig in and figure out your own values. That is the first thing you should be doing. So that's how you invest in mindful change is first by knowing where you're at, understanding yourself. For people who have been in execution mode, finding ways to readopt a learning posture. This is really hard. If you have in your career, if you haven't needed to learn a lot of new things, you know, of course, as engineers, we're constantly learning, right? But if you haven't gone into learning mode, what do I mean by this? I mean, intentionally learning something, not just learning a new command here and there or learning about a new domain because an incident happened and you had to figure that out or learning some technical corner case or something. All of those things are maybe 5% to 10% focus. Learning mode is more like 80%, 90% focus. Your whole posture changes into adopting this new thing and focusing on ways that you can challenge yourself and build new skills as your core. core, your kind of core responsibility or one of your core responsibilities. This is part of investing in mindful change, being able to adopt that learning posture. Some of that comes by making learning a habit, like we mentioned earlier in the show. Rediscovering how your mind responds to change, right, is another part of investing in in mindful change, learning how to relate to a changing world in more useful ways that align with those core values. And if you're able to recognize how you are responding, and especially if you've had an automatic response that doesn't take you where you want to go, You have the automatic response, maybe it's out of fear, maybe it's out of frustration, maybe it's out of fatigue, maybe it's out of habit, mindlessness, who knows, right? If you've had a response that doesn't take you where you want to be, whether that is ambivalent, maybe you're very slow to adopt something, whatever the behavior is that you that doesn't align with your core values, being mindful and trying to find opportunities to change that, right? To learn how to relate to it differently and understand how your adoption pathway here, how are you going to do it? Or in some cases, your exploration is is going to lead you to recognizing that, no, you know what? Actually, I don't think I want to be a part of this. I don't think that this is the direction I want to go. I need to find a new way. It's just as important to recognize that because that is a different kind of mindful change. But that is one way that you can navigate this and in 10 years look back and say that you were proud of how you responded to this major, major upheaval and change in an industry that you worked in. Hopefully this episode didn't throw you off. You know, this episode is very different. And for much of this show, a lot of what I've produced has been guarded a bit. it. It's been, you know, largely I try to avoid putting my opinions too strongly onto the table, except for things that are relatively easy to have a strong opinion about, right? And a lot of times I don't talk about the things that I like, the things that I'm doing, doing mostly because not everybody necessarily needs to hear that all the time. It may not even be helpful to you. I want to focus on the things that are helpful to you. But I've recognized that these discussions more and more are becoming helpful to me. Seeing what other people are doing and connecting with what other people are doing has become a helpful process and a type of media that I think is particularly useful, especially especially right now, as all of our tooling and all of what we've known seems to be up in the air. I just wonder, and I think you do too. I wonder how are people actually going through this? How are you walking through it? I love to hear from you about how this change in the industry, how you're taking it. And I mean that both rhetorically, I want to hear from you, but I also literally do. You can send me an email at developertea at gmail.com. You can also join the Developer Tea Discord at developertea.com slash discord. If you haven't subscribed to this show in a podcast application, whatever podcast app you use, or on YouTube, we now do YouTube videos for every episode. Go and check it out. Head over to developertea.com and you can find the podcast. podcast uh you can find this the show on any podcast player you'll find us on develop at developer t on youtube um thank you again to today's sponsor serp api you can get started for free at serp api.com that's serp api.com thank you so much for listening and until next time enjoy your tea