Building Better Developers with AI

Welcome to Building Better Developers, the developer podcast where we work on getting better step by step professionally and personally. Let's get started. Well hello and welcome back. We are Building Better Developers. We are in a new season. We just finished up our prior one and we are starting a new one. We don't even know what the season topic is going to be yet. We'll talk more on that later. First introductions, I'm Rob Brodhead, one of the founders of Developineur, also a founder of RB Consulting, where we are, as some people call it, a boutique consulting firm where in our case, we sit down and we're very specific about working with our customer to figure out where they're at, what is their business, what is the goals, what are some of the things that are their secret sauce, and then craft a specific recipe for them of using technology and a roadmap to be better in the future, a better business based on leveraging technology. And that includes things like simplification, integration, automation, innovation, whether you're you've got a custom app that needs to be rebuilt, whether you've got something needs to be upgraded, whether you've got too many cooks in the kitchen, you need to integrate some of those systems, give yourself a single dashboard, those kinds of things. We're going to help you find the best solutions out there because we play in that world all the time. And if you don't, so we're going to allow you to keep doing your business, do more to work on your business, and we're going to help you in your business part go a little smoother and a little faster. Good things, bad things. So Memorial Day was just this last week. It was a rainy, rainy day. It was a crappy day. And we were in a little town in Tennessee where basically everything shuts down from Memorial Day and then also was just shut down in general because it was not very good. And by correct, it wasn't like pouring rain, but it was like just a little bit of drizzle. So it's the stuff where you could be out in it. It's just not comfortable to do it for a long period of time. That was the bad thing. Well, actually, part of that was the places we wanted to go for breakfast were all closed. We ended up having it at a fast food restaurant because we just like we wanted some food. We were hungry. Got in, got our food. This is where the good thing is we got out a lot faster than we would have. And so we were just walking through this downtown area and stumbled across a Memorial Day celebration. We got there like just as it started. Like they were just starting the first part of it and got to be a part of, you know, by part of that to witness the entire ceremony and all that kind of stuff. So it was really good to be able to actually embrace Memorial Day in this little town and all of their thanks to the vets and things like that. So sometimes the bad things do lead us to good things. For example, my bad introduction is going to lead you to hopefully a good introduction for Michael. Go ahead. Hey, everyone. My name is Michael Melush. I'm one of the co-founders of Building Better Developers, also known as Developer Nur. I'm also the founder of a company called Envision QA, where we work with businesses, small and large to help you streamline your business. You could be struggling with outdated software. You could be struggling to find the right software to help you run your business. We step in. We work with you. We help you. We walk through the process with you. We understand your processes and procedures. And then we kind of put together that assessment and help you figure out where your pain points are and how to improve your business. And then through the process, we will help you establish what you need to be a more productive and streamlined business so you can make your customers happy and actually hopefully bring in more money. Good thing, bad thing. Well, similar to Rob, Memorial Day was just the past weekend. Even like Rob, I had tsunami-like weather most of the weekend. There were days looking out my window you literally couldn't see five feet past the deck. There was so much rain, it was like Niagara Falls. But the first day, the Friday evening after the rainstorms, it was sunny for about six hours and then three o'clock in the night we had hail, wind, storms. And then it's like, OK, here, we're going to be clear for about four hours. And then boom, here's another tsunami. The good side of that was my pond is back all the way up again. It was kind of going down and my garden is looking nice and healthy. It may be overwatered at this point, only a few days of sun. But the good thing is I don't have to worry about watering my garden or my yard for quite a while. So this season, we talked about a couple of things. So one of the things that we had come across was the idea of maybe integrating AI into each episode so that there would be like maybe a question or something like that that we would go with. And in thinking through it, I think what we want to do this time is we're going to actually use AI potentially to figure out what our episode or season's topic is going to be. So I'm going to walk through this just chat GPT. I want to walk through asking it some questions and we'll see what it says and we'll sort go down some of these rabbit holes. So the first question is, I said, what are good season topics for the developer nor dash building better developers podcast? And the funny thing is, is it gives me looks like eight seasons of topics with about looks like about five to six bullet points for each one. So for example, as some of these I think we've done some of this season one from developer to software craftsman focus is foundational principles and mindset evolution. They talk about clean code and practice, solid principles of real life, tech debt, managing it and avoiding it and writing codes for humans and future you test driven development pros cons when they work. A lot of these sound very familiar. Season two building better software projects, designing software with the focus is project planning and architecture, designing software with scalability, choosing the right tech stack, estimating software projects that don't suck and requirements gathering for developers, agile, scrum, Kanban, what works and why. Season three security for everyday developers. They talk about OOSF and often versus OZ writing secure APIs, secrets management, secure by design season for developer tool belt, mastering the tools, things like get internals every dev should know, setting up CI CD, docker and containers 101, monitoring and logging, using IDs like IDEs like a power user. Season five full stack fundamentals, front end frameworks comparisons, rest versus graph QL, back end with Spring Boot Node or Django, databases, DevOps basics for developers. Season six career development for developers, building a portfolio, certifications, are they worth it? Season seven prep for interviews, avoiding burnout, becoming a tech lead without burning out. Season seven AI and automation for developers. How is AI changing coding, prompt engineering, automating repetitive tasks with scripts, email basics, ethics of AI. Season eight, the last one it gave us, communication and teamwork, writing better documentation, giving and receiving code reviews, navigating tech debt, working with non-technical stakeholders, mentorship and pair programming. And then it says afterwards, it says, well, let me know your show's tone and or audience level, junior devs, team leads, indie hackers, et cetera. And I can help customize these topics further outlaying a full episode structure per season. So let's pause there. What are your thoughts on this? So it's kind of funny because we have actually covered most of that, at least in like the first two seasons. It mentioned the last one, definitely. Some of the other stuff goes back to some of the kind of hands on videos we did with our coding in that back in the day and some of the security stuff we did. It kind of prompted me to want to there's some topics there that would be better served, I think, as going back to our kind of how to videos with kind of walking through software. But the AI one was interesting because the biggest one I hear a lot and I struggle with this one too, is prompting AI to kind of give you what you need. You know, that's one of the biggest struggles with AI, because it's like Dragon Naturally Speaking, you have to train it to understand your way of thinking. And then once you've spent enough time with it, you can pretty much just start with one question and then say, OK, give me more or hey, that's not quite right. Do this. You are training AI, but AI is also training you into how to kind of either if then responses to the conversation that you're having. Exactly. And that's that's part of the problem I've seen in a couple of cases, for example, that I've had more than a few that have gotten mindbendingly circular because there's two essentially there's like two versions of solutions and it'll keep bouncing back and forth between of them and it'll get all the way through one and it'll be like, well, hey, have you tried this? And then you go back to the other and then you're like, wait a minute, which one am I doing? And it'll sometimes mix and match some of those some of the key periods. It's like, make sure you do X, but it'll have it in the wrong stack of steps, essentially. So there's there's definitely some stuff there where you have to train it. You have to do some trial and error and let it know things like, hey, this is what I tried and this is what I'm getting. So those kinds of things. And it will always be like, thank you. I needed that extra context. Here's some other crap. And it really is good for getting you through those the walls that we pound up or, you know, writer's block and things like that. So maybe that's where it sort of points to maybe we'll stick with this. So what do you think? Should we what should we give it for? Let's see. Give a little bit more. So like, what do we want to say the audience level is or maybe the tone? Let's go. So I think I still fairly new. So I would say the audience is going to be AI novice, but probably mid level developer. Or that this is a little on AI. But let's go with this. So I'm going to say the tone is fun and educational. What's a casual? Now, it's interesting, too. Are you using the paid version or the free version? Because there are differences between the two. And we do it. I know that the paid version, I've got like the low end of that. So it's using whatever the chat GPT for four or five, one of the latest version is. Would you get that? I think even free, you just get like very limited amounts of it. And then it has to do with its how fast it will respond as well. So I said this tone is casual, fun and educational audiences and AI novice, but mid level developer. Let's see what it says here. Perfect. With a casual, fun and educational tone and an audience of mid level developers who are AI novices, you can strike a balance between technical depth and beginner friendly explanation. Here's a refined set of season themes tailored. So AI for the rest of us, season one theme demystify AI ML for everyday devs tone. You don't need a PhD to understand this stuff. Episodes. What the heck is AI really? ML versus AI versus data science buzzword bingo. How chat chat, chat GPT like my speaking kind of works, which is a really good little topic. Cool projects to try with AI APIs, AI tools that actually make you faster and some that don't prompt engineering for developers, bias, hallucinations and other weird out quirks, not or weird AI quirks. Weird out a little bit, but I guess. Season two, the developer tool shed. They get back into that. They really just get tips. You'll wish you knew last year. CICD without tears. Dockers containers and WTF is a volume IDA short IDE shortcuts debugging like Sherlock Holmes. So really, it looks like what it's doing here is it's mostly given us. Some moderate, you know, modified topic or titles, but roughly the same kind of subject. So let's see. Would you like help on what outlining one full seasons episodes, scripting an intro or picking guest tools to highlight? So how do we want to let's. What were the first two AI for the rest of us and the developer tool shed? You want to do season one? Let's see what that does. I kind of like the first one that one on season. Because it kind of goes really well with from last season, so because we do talk a lot about developer tools and that, but we also talk a lot about businesses. So with AI, especially we can keep it. Kind of for everyone, be it a developer or business. So absolutely, let's break down season one AI for the rest of us and craft a compelling, casual and educational intro for the season in each episode. The season's goal is to make AI feel approachable, practical and fun for mid level devs. You're just starting to explore it. Season one AI for the rest of us tagline, no math hats, no buzzword fatigue, just real talk on how developers like you can start using AI today. Season intro says be first episode or trailer. Welcome to the developer building better developers. I'm okay. I'm Rob. I'm going to do this like verbatim, so it's going to be AI ish. I'm Rob and this season we're diving into something that's got everyone buzzing, panicking or pretending they already know what it means AI. But don't worry, this isn't some dry academic lecture or enterprise sales pitch. We're calling this season AI for the rest of us. If you're a developer who knows your way around a code base, but has no idea what the difference is between a neural network and a pizza delivery route, this is your season. We'll break down the basics, explore tools you can use right now and talk about how AI can help you not replace you. So fire up your terminal, grab your coffee, let's go. Let's get smarter together. I'll talk about the episode one breakdown real quick. I want to talk a little bit more about what it just gave us. So episode one, what the heck is AI really? Goal, define AI in plain English with relatable analogies. Talking points, AI versus ML versus deep learning. How your app can learn without hard coding logic. Real world examples, autocomplete, Netflix, fraud detection. What AI is not, Skynet, soul-having robots. So that's a pretty good episode to talk about it. Now I want to go back to, this is something I've found really fun with AI in a couple other things that I've done. Yes, it's really good, pretty good for coding. There's a lot of stuff, particularly very repetitive tasks that you can get coded real easy. So the ones that I find that are, I'm always going to have typos or something stupid like that, that'll drag my productivity down, are JavaScript functions and stuff like that. If you want to say, hey, I need to pop a shadow box when this value is above 100, and you give it that script and then it'll give you, here's how you modify your script. And it's pretty darn good. Now sometimes it'll mix a variable or something like that, which is down to the same kinds of stupid debugging that stuff that we will do if we have a typo. But particularly if it's something that's going to take you more than three lines of code, it can be very helpful. There are numerous times that I've had stuff where I know it's going to take, like there's five steps. Each of those steps are going to take five to 10 lines of code. And I can either write all that code and do it myself, or I can take AI, it's going to give me that template for last, lack of a better term. I can then take that, put it into my code, tweak it as I need to, and then I'm going to end up saving myself. Sometimes, you know, an hour or two of coding and just the general walking through is like, how do I solve this problem again? Sort of like having your own code toolbox that we've talked about, your own personal repository. But now it's something that I don't even have to go out there. I can go to AI just as easily. I want to talk though a little bit about this, is what do you think about like that intro? Because I'm finding that it is very, sometimes it writes some really compelling stuff with just a little bit of a tweak. Yeah, I think that was actually a very good kind of summary to come back with for what we essentially could do for this whole season. One thing I will throw out though, because you kind of mentioned using it to get little code snippets. I will throw, this is one of those warnings of using AI, and it's very similar to just surfing the web too to get an answer. Be careful what you ask. So if you're actually looking for command line examples, don't just blindly copy and paste and run. You could crash your entire system. If you are doing things that require more security, more command line, AI is great, but still take their response and go Google it. Look at like maybe Stack Overflow. Look at some reputable sites to get a little more research to confirm that, hey, this is the right approach and not something that's going to take me down or crash my entire network or software. I have had that happen, but I've also had that happen from going to a site and following directions on something from an actual manufacturer site, and they had a typo, and that typo caused things to go awry. So there are pros and cons to everything that we do, and AI is no different. I think with that, it's not just go to, and this is where I think some people get into this loop where they're like, oh, they don't really know it, so they'll go AI, and then they'll go to Google, they'll go backwards and forwards and compare a couple of those. You should make sure that whatever the problem is you're solving, that you understand the problem that you're solving and what the solution should do. Don't just, you know, you should know that, like I said, the example I use, these are the things that you should be able to do on your own because you need to be able to vet what it is that you're given. And definitely take, look very closely because there have been more than a few times that there, yeah, there's a typo or there's something that there's an assumption in whatever the response is that can cause you issues and things like that. So, like I said, use it as like a template or a suggestion or something like that and make sure you understand what it is that you're using. One of the things I've used before is said, hey, go through it and change the variable names or the function names to something that matters, that makes sense that you would use, just to be sure, you know, or just get a snippet of code that's going to require input from other places. And then you're going to have to change the code anyways to handle that input. And that's going to force you to touch it enough as opposed to just like the copy and paste, which I have seen way too often where you see the copy and paste and it's got comments in it that are basically, this is where we do this, this is where we do that, which is very AI-like. And it's exactly what you're going to get. If you throw code at it and say, just comment this code, it's probably going to give you the same kind of thing. So, yeah, as always, take it with a grain of salt. So now what I'm going to do is let's do one more question into this and we'll see because who knows we're going to go from the season still. It says, would you like help writing full episode scripts, interview questions for guests or creating social snippets or summaries for each episode? I'm going to say yes, all three. And so let's see what it comes. Awesome! Let's get you started with scripts, interview questions and social media snippets for season one, episode one, What the Heck is AI Really? So, this works out pretty good. Episode script, outline style, 15 to 20 minutes. That's not bad. Intro, hey everyone, welcome to Develop-a-Noor, Building Better Developers, I'm Rob, and today we're kicking off our season, AI for the Rest of Us. This episode, we're tackling the big question, everybody asks, what the heck is AI really? And of course, we could add to this. This is actually, timing-wise, going to be really good. Part one, defining AI without the buzzwords. And now it's got four bullet points. I'm not going to go too deep, but it actually says, part one, defining AI without buzzwords, three minutes, gives us some bullet points to go through. Classic analogy, if traditional code is like a recipe, AI is like giving someone 100 finished dishes and asking them to reverse engineer the ingredients. That's a really good analogy. Part two, four minutes. AI versus ML versus deep learning. And it basically, here's some bullet points for real world. AI, Roomba that avoids furniture, ML is spam filter that learns from your inbox, DL, deep learning, imagine image classifier that tells cats from dogs. All of which are really good AI solutions. If you haven't looked into them, probably do, because there is a value to each of those for some of the things. For probably any project or application you're doing, you can find where those are going to be useful. So that was three minutes, four minutes, another three minutes, part three, real world examples you already use. Part four, what AI is not, is another three minutes. And then there's an outro of two minutes. Next week, we'll decode the jargon NLP transformers, LLMs, don't worry, it won't hurt. Until then, try spotting how many AI-powered things you interact with this week. Hint, it's probably more than you think. So that's actually, for a summary, that's a really good, I think, approach to how we often do our episodes. Here's some interview questions, optional segment or full episode. Some guest interview questions would be, how would you explain AI to your non-tech family members? What was your first real world AI aha moment as a dev? What are some misconceptions developers have about AI? What's something AI is good at and something it's terrible at? Do you think every developer needs to learn machine learning, or why or why not? Social media snippets for Twitter, X, LinkedIn, threads. New episode drop, what the heck is AI really? Kicks off season one of Develop an OAR, perfect for devs who know code but not transformers. What's the difference between AI, ML, and deep learning? Hint, not just marketing. Season one, episode one breaks it down in plain dev speak. Listen now with a link. And three, if your code is like a recipe, AI is a chef who figures it out just by tasting the dish. This and more metaphors in episode one of our new season, AI for the rest of us. So thoughts on how AI has laid out an episode for us. Yeah, I've played around with AI a lot doing things like that. And it's interesting how concise it kind of got down there, probably because there's been a lot of conversations. I mean, a lot of what we get out of the AI, like chat GPT, copilot, it's what's been fed into, what it's been allowed to kind of scrape from the internet. So I think chat GPT is one of the fewer models that is a little more behind because I think it's, I don't think that one's connected to the internet. So I don't think it can just go out at real time and just go scrape a site. But copilot and some of the other ones can. So I really like some of the ideas and topics it came up with. I really like kind of that first few top ideas because that's kind of what we've just done here, right? We've just kind of walked through a brief overview what AI is. Some of those get a little deep though for kind of, I think the direction of like the season should go. Like I don't think we need to get too deep in the weeds with some of like the higher, like the deep learning and machine learning. Outside of talking about them and giving some good examples, but I think for the first season, we should try to keep it more of that interactive, the higher level AI. So here's a thought I had as I was going through this. One, I just checked, I was in chat GPT and I was like, okay, what's the name of the latest Pope? Because I figured that's newish. And it actually went out to the web and then it gives me, it basically, I guess, does its own effectively Google search or something like that, which is, so it's pretty cool. It gives me some good answers and stuff like that. It doesn't give me, it does give me like a Wikipedia thing, a little bit of a, and it comes from AP News. Oh, and it actually has given me some good, a nice little bibliography here. It's like, you know, here it is, Wikipedia plus 14, Reuters plus seven, a little bit more information, AP News plus two, Wikipedia plus one. Here's something from Vox. So it tells me where the sources are, which is really nice. So, you know, as you go back to those and it gives you some links to each of these. So it does have some usefulness there. That being said, I don't want to digress too much. So here's what I was thinking about. It just sort of came to me for this season. How about we take a past season. Maybe we go two seasons back where we had the challenges, the building better developer, you know, focused on building better developers. And we take each episode and we talk about the topic, but now we're going to go to AI and see what AI thinks about it. That's an interesting. So we just basically repeat the season, the same topics, but we're going to basically bring it, we're going to interview AI for each one of those episodes and say, well, what do you think? And I think it'd be really interesting to see with some of them is maybe we do it first with just like, what does AI think about that topic? And then we take, because we probably can take the transcript. I don't know if it's maybe too big or take part of the transcript of what we did, maybe our blog summary and shove that into AI and say, you know, okay, based on this, what are your thoughts or something like that? So I think like, I think a rehash of that prior season would be a really fun one to see. And it also is going to help. I think it's going to spark some other ideas and things like, maybe new challenges. So thoughts on that? Yeah, that's a lot. That's a lot of fun. And by the way, I believe the transcripts will fit because our episodes are small enough that we shouldn't have a problem doing that. So we have a season topic. This is awesome. We're going to take a prior season topic and we're going to pump it through AI and see what happens. I'm already excited to see how this one goes. I have a feeling it's going to get really interesting really fast. We'll try to, we're going to have to keep it short how much we integrate AI because I have a feeling we're going to have a commentary on every one of its answers, essentially. All right. So we have achieved our goal of a season topic. So this season, we're going to take the prior season, building better developers with AI. So you could think of it with a little superscript like with AI, now with AI or something like that. We're just going to, we're like everybody else. We were jumping on the bandwagon. We're going to do the same thing we did. Now we're going to add AI and see what happens. And I have a feeling this is going to be really fun and educational as well, as far as what AI will give us and what maybe it may, maybe where it could may lead us astray, things like that. Hopefully we won't be starting up Skynet anytime soon based on these questions. That being said, we're going to wrap this one up. As always, love to hear your feedback, your thoughts, info at development.org.com. Shoot us an email, leave us comments wherever you're watching this or listening to this, whether it's out on wherever you listen to podcasts. If you're going somewhere where you don't hear us, where you grab podcasts and we're not there, let us know. This should be, we should be everywhere, but hey, stuff pops up. We may not be there yet. Love to add it there. YouTube, the developer of our channel, we have got, oh gosh, we've got to have close to a hundred episodes of the podcast in video form. And then before that we've got, I'm pretty sure over a hundred, maybe 200 different episodes of how tos of old mentor presentations. There's a lot of content back there. I was working with somebody just the other day and was just talking about like my, the RV consulting site and happened to mention, oh, by the way, there's the developer site and they like their jaw hit the ground with like, that's a lot of content. Like that's a game changing amount of content based on what you were just saying that, well, you know, I've got a few blogs here and a few things. It's like, no, there's a lot of stuff there. So I'm just saying that so that you can go out there and take advantage of it. You know, definitely. And if you have questions about any of that stuff, no matter how old it is, we're happy to feel that update the comment, the content if we need to, or whatever it is to help you become a better developer. Today, go out there and just be a better human as well as a better developer. Go out there and have yourself a great day, a great week. And we will talk to you next time. Thank you for listening to building better developers to develop a newer podcast. You can subscribe on Apple Podcasts, Stitcher, Amazon, anywhere that you can find podcasts. We are there. And remember, just a little bit of effort every day ends up adding into great momentum and great success.