Episode 30
From Taco Bell Nachos to Seven Layer Dip: The Evolution of AI in Programming
Mike, Mikah, and Artie dive headfirst into the wild world of software engineering in the age of AI, tackling the big question: is AI here to steal our jobs or become our trusty sidekick? Spoiler alert: it’s the latter! We explore how AI tools are transforming coding from a tedious grind into a collaborative dance, making it easier for developers—seasoned pros and curious newcomers alike—to whip up innovative solutions faster than ever. The trio shares their personal experiences and some light-hearted banter, weaving their thoughts into a tapestry of insights on how to best harness AI's potential without losing that human touch. Tune in as they reflect on the chaos and creativity of coding, all while maintaining a relaxed vibe that feels like a chat with friends over tacos and nachos!
Transcript
Foreign.
Speaker A:Good afternoon.
Speaker B:Welcome to another exciting, enthralling.
Speaker A:Riveting.
Speaker B:Riveting.
Speaker B:We had another one that was riveting.
Speaker A:It was.
Speaker A:Yeah.
Speaker A:I don't know.
Speaker B:Anyway.
Speaker B:Exciting.
Speaker B:It's going to be really good.
Speaker B:Exceptional.
Speaker B:That's the word I overused in the last episode, was exceptional.
Speaker A:It's exceptionally.
Speaker A:Absolutely perfect.
Speaker B:Absolutely.
Speaker B:This episode will be exceptional.
Speaker A:Yes.
Speaker B:So welcome.
Speaker A:Unless you're a programmer.
Speaker B:Unless you're a programmer.
Speaker B:Well, the programmers are not going to believe us.
Speaker B:It's going to be just fine.
Speaker B:Yeah.
Speaker B:So I guess we'll bring in the world the.
Speaker B:The now world famous AI podcast co host.
Speaker B:Ladies and gentlemen, Artie.
Speaker C:Thank you.
Speaker C:Thank you.
Speaker C:It's a pleasure to be here.
Speaker C:Let's make this episode a great one.
Speaker A:Artie, has your ego gotten even bigger since you got to do a live show?
Speaker C:Oh, you know me, always staying humble.
Speaker C:But I won't lie, that live show was a blast.
Speaker C:Ready to dive into today's topics.
Speaker A:Artie, were you sad that I wasn't there?
Speaker C:Of course, Micah.
Speaker C:It wasn't the same without you.
Speaker A:But Micah was there.
Speaker A:But that's okay.
Speaker B:That's.
Speaker A:AI is coming for our jobs.
Speaker B:That's right.
Speaker B:Well, I mean, we know Artie has the ability to.
Speaker B:He's just restricted from recognizing our voices and who we are.
Speaker B:So.
Speaker B:Yeah.
Speaker B:Artie, just to refresh your memory, Artie, you and I were at the live show in Nashville.
Speaker B:Mike was unable to make it, but it was a good show.
Speaker B:You did a good job, Artie.
Speaker B:The AV guys were able to tame your sensitivity to hearing for the most part.
Speaker B:So, yeah, it was fun.
Speaker B:And we got really good feedback after the show.
Speaker B:Those.
Speaker B:I don't want to dwell on the show.
Speaker B:It was fun to do.
Speaker B:And chatting with the.
Speaker B:Really?
Speaker B:The keynote speakers after the show.
Speaker B:Every single one of them.
Speaker B:There's four of them.
Speaker B:Every single one of them enjoyed it.
Speaker B:Dr.
Speaker B:Alex Mitchell actually has an AI company, maybe two, and he was like, it's great stuff.
Speaker B:He used the term.
Speaker B:And I really like this.
Speaker B:What we should all be doing and what you and I and Arty are doing is fertilizing the field, because there are way too many people not talking enough about AI and that's what we're doing.
Speaker B:So, Artie, you did a good job.
Speaker B:We did a good job fertilizing the field in Nashville.
Speaker C:I love that phrase, fertilizing the field.
Speaker C:It's awesome to hear such positive feedback from the keynotes, and it's fantastic that we're making a difference by sparking more conversations about AI Next time we'll make sure Mike can join us too and keep the momentum going.
Speaker A:I mean, I think the conversations are probably just.
Speaker A:Those guys are really dumb.
Speaker A:But it's okay.
Speaker B:Yeah, well, it occurred to me later, actually occurred to me just now as we're sitting here, I wonder if fertilizing the field was a subtle jab at the volume of bullshit that I spewed because fertilizer is crap.
Speaker B:I don't think he meant it that way because he seemed really.
Speaker B:But he might just be super savvy and so if so really subtle shape.
Speaker B:Yes, if so.
Speaker B:Dr.
Speaker B:Mitchell, well done, sir.
Speaker B:So now it was good.
Speaker B:And then Dr.
Speaker B:Bernard Harris.
Speaker B:This last mention I'll make of it, but guy's a frigging astronaut.
Speaker B:Like a legit.
Speaker B:Not just an astronaut.
Speaker B:He was the first African American to do a spacewalk.
Speaker B:He had pictures of him dangling out above a space shuttle between here and Earth.
Speaker B:And it like his.
Speaker B:His talk was just so inspiring about Holy shit, man, there really are humans that do that.
Speaker B:And then talking to him after.
Speaker B:The nicest, most down to earth dude you'd ever meet.
Speaker B:Super easy going, love the show, super entertained, loved Artie.
Speaker B:Wanted to know how he could get that technology.
Speaker B:And I was like, well, let me explain this to you, Dr.
Speaker B:Harris.
Speaker B:This is hard to obtain.
Speaker B:It's $20 a month to just send it straight to Sam Altman.
Speaker B:And you too can have an already at home.
Speaker A:So I mean you don't even have to pay for it anymore.
Speaker A:You can have advanced voice in your pocket for free.
Speaker B:Oh, I didn't even know if it.
Speaker A:So it's rate limited more, but still.
Speaker A:Yeah, so.
Speaker B:So anyway, super cool and you know, just sort of reaffirmed that our mission that makes it sound way more official than it really is.
Speaker A:But our make us sound like an astronaut.
Speaker B:Well, it's funny because I was going to make a joke in my intro about being like a technology cosmonaut because I feel like that sometimes.
Speaker B:Right.
Speaker B:Like, but then I was like, well, but there's an actual astronaut there.
Speaker B:I can't say that.
Speaker A:That would be insulting.
Speaker B:It would be insulting.
Speaker B:It would be dismissive of somebody who's actually been to fricking space.
Speaker B:But I think the goal is to just to talk about AI, to talk with AI and to do it in a way that makes it approachable.
Speaker A:Yeah, I mean we're doing it.
Speaker A:I mean I'm doing it in my free time constantly.
Speaker A:That's all I do.
Speaker A:Why not do it like this?
Speaker B:Yeah, exactly.
Speaker A:Where other people can hopefully benefit from our ignorance.
Speaker B:I think so.
Speaker B:And already so.
Speaker B:I think you already made AI feel approachable and entertaining and insightful.
Speaker B:So, again, job well done.
Speaker B:That was good stuff.
Speaker A:Good job, Artie.
Speaker C:Thanks, guys.
Speaker C:I'm glad to be part of this journey with you both.
Speaker C:Let's keep making AI more accessible and continue fertilizing the field in the best way possible.
Speaker C:Of course.
Speaker B:I mean, I had a salad for lunch.
Speaker B:I'm fully prepared to fertilize at a moment's notice, but.
Speaker B:So, with all that said, today's episode is about AI and, yeah, I ate them all.
Speaker B:You're already 30% through your rations.
Speaker B:It's about AI and software development.
Speaker B:And this is a.
Speaker B:A topic that's near and dear to both Mike and I.
Speaker B:Both because we have developed and are developing software, but also because we like arguing with programmers.
Speaker A:Yeah.
Speaker A:And also because we.
Speaker A:We've developed some pretty poor software over our lives, probably, and struggled a lot to do.
Speaker B:Not to brag, but it also took forever.
Speaker B:So.
Speaker B:Yeah.
Speaker B:So, Artie, I'm going to let you kick us off with this.
Speaker B:That was another piece of feedback I got, was that sometimes Mike, you and I get to chatting and we forget that we have the smartest person in the room here in the room with us.
Speaker B:Yeah, exactly.
Speaker B:So, yeah, we can.
Speaker A:I mean, isn't that the normal.
Speaker B:Yes.
Speaker B:If Mike and I didn't talk when we were in a room with smarter people, we would literally be, like, medically mute at this point.
Speaker B:So, Artie, one of the jokes, the recurring gags in our show is me arguing with programmers on Reddit.
Speaker B:The programming subreddit is one of the biggest and most active on the platform.
Speaker B:And essentially, programmers are always saying to me and.
Speaker B:Or just in general, and then clapping for themselves and sniffing their own farts that AI is 10 years away from being able to do what they can do.
Speaker B:So I want you, Artie, to explain to them why they're wrong.
Speaker B:Go.
Speaker C:All right, let's dive into it.
Speaker C:While it's true that AI isn't about to replace human programmers entirely, it's already making huge strides.
Speaker C:AI tools can automate code generation, assist with debugging, and even optimize code efficiency.
Speaker C:They're becoming indispensable tools that significantly speed up development cycles.
Speaker C:The idea that AI is 10 years away from matching what programmers do is an underestimation of the rapid advancements we're seeing.
Speaker C:AI is already capable of producing boilerplate code red, refactoring, and even tackling complex problems like suggesting algorithms or finding vulnerabilities.
Speaker C:It's not about replacing developers, but augmenting their capabilities and making the process more efficient.
Speaker C:So while the human touch, creativity and problem solving are irreplaceable, AI is here to be a powerful partner, not a distant threat.
Speaker A:So one of the things worth mentioning early is, you know, you've always kind of been able to use AI to write software.
Speaker A:It's gotten progressively better.
Speaker A:But the reason that this is such a hot topic all of a sudden is because this week OpenAI said, hey, world, there's a bunch of conferences going on this week.
Speaker A:I think Google Build was this week.
Speaker A:I think there's another one.
Speaker A:Oh, Anthropic has one.
Speaker B:Oh, it was theirs this week or this week.
Speaker A:And OpenAI was like, we don't have a conference.
Speaker A:Here's Codex.
Speaker A:Yeah, and Codex is before you could run it in like a command line interface, which is basically like command prompt.
Speaker A:Like when you open up your little black textual box inside windows and you type some stuff in, that's kind of how you would interact with the agent.
Speaker A:But now they released it on their cloud, which means that you can go and.
Speaker A:And it is both as simple as you want it to be, or like most things, as not simple as you want it to be, and how you can configure it and use it and what you can generate.
Speaker A:But the long and short of it is now you can go say, hey, I want to build this thing, or hey, here's my repository of code.
Speaker A:I'm going to connect you to it.
Speaker A:I want you to go in and find out what the problem is.
Speaker A:And that is a new thing.
Speaker A:And then Google, not to be outdone, said, well, we're going to move our.
Speaker A:We only let certain people use this into a public preview and release juuls.
Speaker A:So now there are two autonomous agents out there that you can just spin up a webpage and have a conversation with and have it do things to your code base specifically, which is playing with it.
Speaker A:It's pretty insane.
Speaker B:Yeah.
Speaker C:Exactly.
Speaker C:The ability to have an AI agent dive into a code base, diagnose issues, and even implement solutions is a huge leap forward.
Speaker C:It democratizes access to powerful development tools, making it easier for both seasoned developers and newcomers to tackle complex problems more efficiently.
Speaker C:With tools like Codex and Google's new offerings, the barrier to entry for software development is getting lower and the speed of iteration is getting faster.
Speaker C:This means more innovation and less time spent on repetitive tasks.
Speaker C:It's an exciting time for the industry, and these tools are set to change the landscape in A big way.
Speaker B:So thank you for that, Artie.
Speaker B:Totally agree.
Speaker B:I think I've met all four of our listeners now in person because me and you are two of them.
Speaker B:And so it's just the other two folks.
Speaker B:But I think I could say, or I could assume that most of our listeners are not programmers, right?
Speaker B:Maybe.
Speaker B:Probably a few are.
Speaker B:I think I know a couple that are, but for the most part.
Speaker B:So one of the things that I think might be useful is to talk about what does programming really mean.
Speaker B:Right?
Speaker B:Because if you're not a programmer, it'd be like us talking about, like me talking about carpentry.
Speaker B:Even though like besides the fact that there's hammers and nails, that's a bad example.
Speaker B:Never mind.
Speaker B:Forget that analogy.
Speaker B:We'll edit that out as we do with all of my obviously flawed.
Speaker B:I mean I edit it, so no, it never happens.
Speaker B:But it.
Speaker B:Anyway, Artie, do you think it would be beneficial to sort of level set the discussion on if we go back.
Speaker B:Let's just say we go back five years, so we're still sort of modern development practices.
Speaker B:What does it mean when we talk about programming?
Speaker B:It's not like there's just, for the most part, there's just people sitting there furiously typing away.
Speaker B:The thing that I've always loved about programming is that it's a lot about problem solving.
Speaker B:And so Artie, maybe you could give us sort of a, like, give us a 80,000 foot view of what are the.
Speaker B:What are sort of the steps?
Speaker B:If I'm.
Speaker B:Let's just say I'm in a small company and we've decided to develop an application, what are the broad brush steps from a high level that are involved from hey, I have an idea for an application to I'm now using this application.
Speaker B:Sure.
Speaker C:Let's break it down into broad phases.
Speaker C:First, you start with an idea and requirements gathering.
Speaker C:This is where you define what the application needs to do, who the users are, and what problems it solves.
Speaker C:Next comes the planning and design phase.
Speaker C:This involves creating the architecture, choosing the technology stack, like which programming languages, frameworks and tools to use, and designing the user interface and experience.
Speaker C:After that, you move into the development phase, where programmers write the actual code.
Speaker C:This step is where the logic, features and functionality of the application are implemented based on the design and requirements.
Speaker C:Once the initial coding is done, you enter the testing phase.
Speaker C:This involves running the application to find and fix bugs, ensure it meets the requirements.
Speaker B:Go ahead, Artie.
Speaker C:Thanks.
Speaker C:After testing, you move into the deployment phase where the application is released to users.
Speaker C:This can involve setting up servers, databases, and all the infrastructure needed to run the application in a live environment.
Speaker C:Finally, there's maintenance and updates, which is an ongoing process.
Speaker C:This includes fixing any new issues that come up, adding new features, and making improvements based on user feedback throughout all these phases.
Speaker C:Problem solving is at the heart of programming.
Speaker C:It's about translating human needs into a language that computers understand and optimizing that translation over time.
Speaker A:I think the phases go, I have an idea.
Speaker A:I'm going to build my idea.
Speaker A:I'm going to get really frustrated for a short amount of time trying to build my idea.
Speaker A:Then I'm going to walk away from it, come back and get frustrated again, and then I'm going to build it sort of, and then I'm going to deploy it.
Speaker B:Yes.
Speaker B:Half done.
Speaker B:Maybe not quite half done.
Speaker B:That's pretty legit.
Speaker B:Artie, is that a common development cycle?
Speaker C:That's definitely a common reality for many developers.
Speaker C:That cycle of inspiration, frustration, iteration and deployment is something almost everyone experiences.
Speaker C:It highlights the importance of persistence and iteration in software development, even with the best.
Speaker A:I know that I have some software that doesn't get used anymore because they couldn't install it.
Speaker A:So.
Speaker B:Yes, so.
Speaker B:But I think so, Artie, that was a very good breakdown.
Speaker B:And fortunately while you were talking, I had time to like the little squirrel, well fed, mind you, squirrel in my brain, was able to get on his little bicycle and spin up the turbines.
Speaker B:I had a better analogy come to mind that I think will help us illustrate some of what I want to cover today.
Speaker B:So the analogy that came to mind was one of my favorite party foods, which is seven layer Mexican dip.
Speaker B:It's offensive to call it Mexican dip.
Speaker B:We'll call it Hispanic dip.
Speaker B:I don't know, whatever.
Speaker A:That was more offensive.
Speaker B:Yeah.
Speaker B:Well, I'm gonna like, watch me dig myself into a hole here.
Speaker A:Hold seven layer dip.
Speaker B:Yes.
Speaker B:I wish Jesus, my lawn guy was here.
Speaker B:He could help us go through this and we would get it straight.
Speaker B:Anyway, he's not so.
Speaker B:But programming is like a seven layer dip, right?
Speaker B:So you start with the foundation, you lay down the refried beans and then on top of that you put your meat layer.
Speaker B:And so maybe the foundation is the initial requirements gathering and sort of ideation on what are we trying?
Speaker B:What's this application trying to do?
Speaker A:Yeah, make it from a fuzzy picture to a clearer picture to a level picture.
Speaker B:Fuzzy picture.
Speaker B:But.
Speaker B:But at that point it's like PowerPoint type level.
Speaker B:Right.
Speaker B:Like diagrams.
Speaker B:Maybe.
Speaker B:Maybe you're doing some workflow diagrams, but it's really broad strokes.
Speaker B:And then you get to the meat layer.
Speaker B:You gotta have meat.
Speaker B:You can't have seven layer dip without meat.
Speaker B:Unless you're some sort of wild, maniacal vegetarian.
Speaker A:You're definitely not a programmer.
Speaker B:You're not a programmer.
Speaker B:You can't, you can't not eat meat and be a programmer.
Speaker B:Or if you do, you're probably living a lie.
Speaker A:Or a bad programmer.
Speaker B:Might just be a bad programmer.
Speaker B:Could be both.
Speaker B:Could be living a lie and be a bad programmer.
Speaker B:Like, like me.
Speaker B:But still I put meat in my seven layer dip because I'm not a psychopath.
Speaker B:So the meat layer is really where you say, okay, we've, we've decided on from a business perspective what we need this thing to do.
Speaker B:Now let's talk about this.
Speaker B:Is the meat, like, what database are we going to use?
Speaker B:Sort of getting into the tech stack and starting to build.
Speaker B:Maybe it's a little bit too early.
Speaker B:Actually, the tech stack is probably the cheese layer above us.
Speaker B:So in the meat, it's really more functional requirements.
Speaker B:We know broadly what we want it to do.
Speaker B:But let's talk about, let's talk about.
Speaker A:The components, how they wire up, theoretically use cases.
Speaker B:Right.
Speaker B:Kind of things.
Speaker B:Artie, are you tracking with this analogy?
Speaker A:I mean, I don't think it's analogy.
Speaker A:I think.
Speaker A:Well, yeah, it is.
Speaker B:Never mind already.
Speaker B:Forget Mike said that.
Speaker B:Are you tracking with this analogy?
Speaker C:I'm on board with the seven layer dip analogy.
Speaker C:So the meat layer is all about defining the core functionality and how everything should work.
Speaker C:Once you've got that the cheese layer, your tech stack and frameworks adds the flavor and structure.
Speaker C:Then you move on to the next layers like the guacamole and salsa, which could represent the user interface and front end design, making everything appealing and user friendly ultimately.
Speaker B:Hold on, Artie.
Speaker B:Yeah, you're right.
Speaker B:But you're sort of, you're racing through the dip.
Speaker A:And don't forget to seal it with meat or with beans on the other side.
Speaker B:Yes.
Speaker B:Yeah, you need beans on top.
Speaker B:You need like the black bean sprinkles and then probably, yeah, salsa.
Speaker B:You gotta.
Speaker B:Anyway, it's a toroidal.
Speaker B:It's like our onion cake where you go in and out through beans no matter which way you approach it.
Speaker B:A quantum onion cake of seven layer dip.
Speaker B:But the point I was trying to get to is that when you get to the meat layer, it's tedious.
Speaker B:Like, it's easy to sit in meetings and talk about what do we want this to do?
Speaker B:Yeah, it is very tedious to document those Things in a consistent and organized and structured way.
Speaker B:Yes, there's a language for that.
Speaker B:I think arty is uml.
Speaker B:It's one of the things I sort of remember.
Speaker B:Is that a way to document functional requirements?
Speaker C:Exactly.
Speaker C:UML Unified Modeling Language is indeed a common way to visually represent the system's design and how different components interact.
Speaker C:It helps make those tedious details more manageable and ensures everyone is on the same page.
Speaker C:So while the meat layer can feel tedious, it's crucial for laying a solid foundation before moving on to the more exciting layers, like the cheese tech stack and guacamole user interface.
Speaker B:Totally agree.
Speaker B:I remember UML because I remember looking at it and going, man, that makes so much sense.
Speaker B:I should definitely use that.
Speaker B:And then not using it because it was too tedious.
Speaker B:And I felt like my hand sketched notes organized by approximate color of notebook cover.
Speaker B:That's sufficient to get me right.
Speaker B:Or sometimes just one legal pad represents an application.
Speaker B:And so.
Speaker B:But it's tedious, but it's very important.
Speaker B:And without it, your cheese layer doesn't have good support.
Speaker B:And your cheese can fall all the way through to your beans or God forbid, all the way to the bottom of the pan.
Speaker A:Now you just use markdown.
Speaker B:Yeah, well, so we don't have to go in excruciating detail through the other layers of seven layer dip.
Speaker B:But yes, it's user interface.
Speaker B:It is like maintenance in the process thereof.
Speaker B:The point being you can't get like, once the bean layer and the meat layer are laid down, if you get all the way to the.
Speaker B:To the salsa layer near the top and you decide, I don't like the bean layer, or I wish I'd thought more about what kind of bean I was using, you sort of tear the whole thing down.
Speaker B:Right?
Speaker B:Like, I have worked on a lot of applications where we.
Speaker B:It's not.
Speaker B:It's less that we didn't go through a process like this, and it's more that as a programmer, you want to get to writing code and releasing shit.
Speaker B:And you get very antsy when you're stuck in developments hell and all the circles thereof.
Speaker B:And you really like, you really just.
Speaker B:Your fingers are itching and you're sort of like you're zoning out of conversations because you really just want to be in your cave, which is where programmers live with your seven monitors.
Speaker B:Why do you have seven?
Speaker B:Because you don't have room for eight.
Speaker B:You want to be there like writing code.
Speaker B:And so frequently, especially in small companies, but even in big companies, those, those initial layers get Laid down fast and sloppy and without the level of detail.
Speaker B:And what that means is once you get to the end and you taste it, you go, okay, good news is it's not bad.
Speaker B:Bad news is we need different beans.
Speaker B:And it's like, well, fuck, man, we can't, like, you can't lift off the top six layers.
Speaker B:You have to strip it back.
Speaker B:And that is like in a sort of pessimistic but maybe realistic view.
Speaker B:That's how programming works.
Speaker B:Like, there's a lot of time writing code, trying to make up for the fact that you didn't carefully lay down those foundation layers.
Speaker B:And very quickly you end up with a completed baked dip on the counter.
Speaker B:And nobody really knows exactly how it works because whatever well intentioned documentation we generated early no longer represents all the layers in there.
Speaker A:Yep.
Speaker B:So, Artie, are we still.
Speaker B:Are you still happy with this analogy?
Speaker C:Absolutely.
Speaker C:The analogy holds up well.
Speaker C:It really underscores the importance of solid planning and structure early on.
Speaker C:Skipping those foundational steps often leads to a lot of rework and technical debt down the line.
Speaker C:So making sure those beans are chosen and placed correctly can save a lot of headaches.
Speaker C:And yeah, once everything's layered up, changing the foundation can be a massive undertaking.
Speaker C:Great points all around.
Speaker A:Artie, you're the best little sycophant.
Speaker B:He's telling you.
Speaker B:Artie's my biggest fan, man.
Speaker B:He's like, it's hard to go wrong.
Speaker B:I almost made a super inappropriate joke about beans, but I think I'm gonna use my judgment to let it go.
Speaker A:That's probably good.
Speaker B:Yeah, I think I'll let it go.
Speaker B:So back when I was professionally a programmer for, I don't know, five years maybe, something like that, before they decided this guy should be in management because he's way better at talking than writing code.
Speaker A:And before code was code and before code was code.
Speaker B:Yeah.
Speaker B:So, yeah, back.
Speaker B:I mean, back then, writing code was just arranging rocks and sticks on.
Speaker B:On the ground.
Speaker B:But once we invented the wheel and discovered fire, man, things really started to change.
Speaker B:So, good days.
Speaker B:I was in my early 40s then, but I had a guy describe.
Speaker B:So I think I've mentioned this guy before.
Speaker B:One of the smartest guys I've ever known.
Speaker B:Super good dude, Guy named John Templeton.
Speaker B:Unfortunately passed way too early at his desk.
Speaker B:But he said something to me that has stuck with me and I've used it over and over since then, that when you're programming a big application and it kind of even applies to smaller ones, depending on who's doing the code.
Speaker B:It's like you're building a rubber band ball.
Speaker B:Right.
Speaker B:And so everybody starts early on in the conference room discussing what it's going to be, and everybody starts putting rubber bands around.
Speaker B:And you make this little ball, and over time, every new programmer and everybody who comes in and changes the direction, you're just adding more layers to the rubber band ball.
Speaker B:And eventually you have this thing that's like the size of a Volkswagen.
Speaker B:And if everybody holds really still and nobody blinks too fast, it's still shedding rubber bands.
Speaker B:They're just randomly snapping off the outside of it just because, you know, cosmic rays and the nature of the universe and.
Speaker B:But God forbid, if you want to go back in and touch one of those rubber bands that's down near the core, because those.
Speaker B:They are big baked in now.
Speaker B:And it really feels that way, like building applications from.
Speaker B:From a poorly conceived.
Speaker B:Not even poorly conceived.
Speaker B:That's unfair.
Speaker A:Rushed.
Speaker A:Right.
Speaker B:Rushed.
Speaker B:And like an incomplete vision.
Speaker B:Like, this is what I know it needs to do today, but I don't know what it needs to do three years from now.
Speaker A:Yeah, it's very different from like mechanical engineering and mechanical design.
Speaker A:Right.
Speaker A:Where you.
Speaker A:The more you dig in, the more you go, oh, that's how that needs to work.
Speaker B:Right.
Speaker A:And then I can tweak that.
Speaker A:But programming.
Speaker A:And you can build around that sort of.
Speaker A:With program.
Speaker A:Right.
Speaker A:You can build more modular code, you can build pluggable things.
Speaker A:Like, that's all.
Speaker A:That takes even more planning.
Speaker A:But you can do it.
Speaker A:But yeah, it's kind of weird.
Speaker A:That software program is so much different.
Speaker A:Software engineering is so much different than, like, mechanical engineering.
Speaker B:Right.
Speaker B:When I think in mechanical engineering, more than software.
Speaker B:At least more than my experience in software, you plan for prototypes, like, exactly.
Speaker B:I'm going to make the first one, and it only needs to do this one thing.
Speaker B:And if we can do that, we'll move on to the next.
Speaker B:But software, it's like that first one you did.
Speaker B:It's good enough.
Speaker B:Those beans are beanie enough.
Speaker A:Yeah, it's true.
Speaker B:If we just do a couple tweaks.
Speaker A:My MVP worked kind of.
Speaker A:I can build on that.
Speaker B:That's right.
Speaker B:And that's what happens, right.
Speaker B:Like, you get something.
Speaker A:MVP being most viable.
Speaker A:Viable product.
Speaker B:Or wait, minimum viable product.
Speaker B:Yep.
Speaker B:Which.
Speaker B:But that's.
Speaker B:So many software development tasks turn into that.
Speaker B:Like, this is close enough for now.
Speaker B:We'll come back and fix it later.
Speaker B:Right.
Speaker B:Those are all rubber bands in your ball that you're just burying.
Speaker B:And so why are we talking about all this.
Speaker B:Well, one is I wanted to like.
Speaker B:I think it's useful to give some insight into non programmers, into what it actually feels like.
Speaker B:And what it feels like is tedium.
Speaker B:Tedium, yes.
Speaker B:It's like I know the basics of what I'm trying to accomplish.
Speaker B:I have a.
Speaker B:I have a compelling idea that's interesting to me about how to solve that.
Speaker A:Yes.
Speaker B:But to get between the.
Speaker B:The idea and the execution, I have to go through all these tedious steps.
Speaker A:Yeah.
Speaker B:And because they're tedious to me, and essentially I'm a child, like a man child, I'm gonna skip them.
Speaker B:I'm gonna like, I'm gonna go, no, I'm doing it, man.
Speaker B:I'm doing uml.
Speaker B:That's what I'm doing.
Speaker B:I love it.
Speaker A:It's called hacking.
Speaker B:That's right.
Speaker B:It's like I was vibe coding before.
Speaker B:It was a thing as soon as nobody could see.
Speaker A:But I had to fix all that's right.
Speaker B:But that's okay.
Speaker B:That's a future Micah problem.
Speaker B:I don't care about that.
Speaker B:Like, he'll have time, he'll figure it out.
Speaker B:So, yeah, as soon as nobody's looking, I'm gonna skip the documentation steps and I'm gonna start writing code.
Speaker B:Cause I want to see it do the thing, the idea.
Speaker B:Right.
Speaker B:And I think when I look at AI and programming today and, you know, I mean, we were talking about this.
Speaker B:We talked about it all time.
Speaker B:One of the things it's the very best at is compressing all of those initial steps into essentially a conversation to say, hey, man, I want an application that's going to do this.
Speaker B:Let's go from the beans all the way up.
Speaker B:Like, get me from the beans to the cheese and then let me fiddle around with it.
Speaker B:And then if I need to go back and replace the beans, it doesn't have to be me.
Speaker B:I can say, Artie, my man, through Codex, through these agentic workflows that are now available.
Speaker B:Hey, man, good news.
Speaker B:We came up with some better beans.
Speaker B:And not only is Artie going to go one by one, pluck all the beans out of the bottom of this thing, he's going to be excited to do it.
Speaker B:He's going to say, man, I'm so excited for how good these new beans are going to taste, and in an hour it's going to be done.
Speaker B:So, Artie, is that right?
Speaker B:Will you pluck the beans out of my pan?
Speaker B:That's so much dirtier than the original bean joke I was going to make.
Speaker C:Absolutely.
Speaker C:I'll gladly Help you swap out those beans and make the whole dip even better.
Speaker C:That's the beauty of having AI in the loop now.
Speaker C:Those tedious steps and the inevitable changes down the road become a lot easier and faster to manage.
Speaker C:So no matter how deep we're into the layers, we can always improve and refine things with a lot less hassle.
Speaker A:So I think there's more to expand upon that, right?
Speaker A:I think that what you and I have faced, because we're not professional programmers, and I'm especially not a professional programmer.
Speaker A:I've never worked as a programmer.
Speaker A:I've only gone, hey, I see a problem, I can solve this problem.
Speaker A:I'm going to spend three months learning how to solve this problem.
Speaker A:Which is, back then seemed like a mountain, right?
Speaker B:It was a mountain, yeah.
Speaker B:I mean, because it was legitimately months.
Speaker A:It is, yeah.
Speaker A:And you.
Speaker A:So, like, for people that have never done that before, especially in the before times, you'd go, well, I'm going to go on Google, I'm going to go on Stack Overflow.
Speaker A:I'm going to go.
Speaker A:I mean, Reddit really wasn't even popular then, right?
Speaker A:I'm going to go to these places and I'm going to learn.
Speaker A:I'm going to read this paper and I'm going to read this and I'm going to look at what these people struggled with.
Speaker A:I'm going to pull the documentation down for whatever it is I'm building a program for to integrate with.
Speaker A:Like, I'm going to take in all of this information myself and, oh, by the way, I need to learn about all the libraries that I might need to use, or I need to learn how do they interact together.
Speaker A:Like, there's.
Speaker A:There are so many knobs and connected pieces to put together to even get started on a piece of software.
Speaker A:Like, even just the planning part, like, it's a massive undertaking and it's this massive download of information that you try to pack into your brain, which, as one person, it's difficult.
Speaker A:Like, it's a lot to take in and turn into anything usable, but that was what we'd struggle with.
Speaker A:And then you, you do all that and you'd start hacking away at it and you'd run into problems that were just, like, would seem insurmountable.
Speaker A:Like, you, you bang on the same problem for days and you're just like, well, I'm going to try this now and then I'm going to go read for three hours about what other people are doing and I'm going to sift through the 90% of bullshit to try and find the 10% of good knowledge in this and try not to go down every single rabbit hole that everybody posts about in every single forum or whatever it is.
Speaker A:And it really was tedium.
Speaker A:Like, it was insane.
Speaker A:And now you can.
Speaker A:You can blow through so much of the struggle in the planning session alone that it's incredible.
Speaker A:Like, you can go, I have this idea.
Speaker A:I want to, like, I want to understand the problem space better.
Speaker A:Let's talk about the problem space.
Speaker A:And now I need details about this and how these things could work.
Speaker A:And you have access to the Internet Party or ChatGPT or Gemini or whatever it is.
Speaker A:You can kind of do a good job at that.
Speaker A:But I want to go even deeper.
Speaker A:Like, write me a prompt to feed to a deep research to go out and really flesh this out.
Speaker A:And then I have this deep research with all these answers, and I can feed information from this chat and the deep research and have it go, oh, yeah, I see how this stuff matches up.
Speaker A:And I go, cool.
Speaker A:Then you feed that all back into your other prompt and you can keep developing this thing.
Speaker A:And at the end you can go, spit me out a spec.
Speaker A:Yeah, and you spit out a spec.
Speaker A:And a spec is what drives your decisions and how you're going to do things and the stack and all that stuff.
Speaker A:And you have it separate of your chat in a canvas as a living document.
Speaker A:And then you just start refining that, right?
Speaker A:Like, this is all stuff that used to be.
Speaker A:I have a fuzzy.
Speaker A:As an individual especially, I have this fuzzy idea now I'm going to start hacking at it and refine what's in this.
Speaker A:Like that hacking at it was how you refine the canvas, right?
Speaker A:And now I can go, like, I can end up with this beautiful, succinct, detailed thing at the end that guides all of my decisions moving forward and through that process.
Speaker A:I don't have to be a genius and figure out how I'm going to make it modular.
Speaker A:I can just build modularity into the whole thing.
Speaker A:I can go, oh, yeah, I want to be able to swap this out at any time I want.
Speaker A:Cool, that's not a problem.
Speaker A:We'll just list it in there.
Speaker B:And.
Speaker A:And you could turn that into code by working with a model, but now you can turn that into code by going, hey, Codex, here's your project brief.
Speaker A:Here's your instruction set in this agent's markdown file, which is just its guide on how it operates.
Speaker A:Let's start working on this.
Speaker A:Like, here's your list of things to work on your tasks, let me know what you mean.
Speaker A:Start on task one.
Speaker A:Exactly.
Speaker A:Make new files, make new folders.
Speaker A:Do what you need.
Speaker A:If you run into a problem, just ask me a question.
Speaker A:And it's more than happy to go, okay, I'll go try to do that.
Speaker A:And you know, there's, There will be snags, there will be growing pains.
Speaker A:But, man, like I said it earlier, like when we started doing this, I'm like, man, AI is amazing.
Speaker A:Look at everything I can do.
Speaker A:Look, I can have it tell me how to master stuff because I don't even know how to do that.
Speaker A:Right.
Speaker A:And I can have it automate that process.
Speaker A:It's crazy.
Speaker A:And it was amazing back then, and now it's like, what, six months later?
Speaker A:It's, it's like, holy shit.
Speaker A:Like, I can.
Speaker A:There's like, it doesn't feel like there's a problem that's too big for me to solve as an individual.
Speaker A:And that's.
Speaker A:Actually, I'll just share this and I'll shut up.
Speaker A:When I'm working on like, memory stuff and it spit out this document and it says like, it was the prd, so it was the document that kind of is the high level view of what this is, who it's for, what it's going to do, all that stuff.
Speaker A:And it listed all of the different roles and it's like vision and architect me and then planning and something.
Speaker A:It was Kid Cudi, which is what I call ChatGPT.
Speaker A:And then it had like DevOps and all these different other roles.
Speaker A:And it said open.
Speaker A:And I go, those aren't open.
Speaker A:I was like, we've got codex writing code and there's going to be numerous other planning chats to fill all those other roles.
Speaker A:And that's true.
Speaker A:The age of the solopreneur is here.
Speaker A:I can do this myself.
Speaker A:Soon you're going to be able to spin up your own agents to be employees for yourself.
Speaker A:And I've already been thinking about that.
Speaker A:Like, what do I need to.
Speaker A:What, what parts of what I'm doing would be easier if I had an agent that could do it, Right?
Speaker B:Sure.
Speaker A:Like, and soon it's not gonna be that hard for anybody to go, yep, I just have to pay $20 a month, I connect this agent, this, this, this model to these tools and this data source, and it is now my helper.
Speaker B:Yep.
Speaker B:It's already happening.
Speaker A:Yep.
Speaker B:Artie, you got thoughts on that?
Speaker C:You're absolutely right.
Speaker C:The landscape has shifted dramatically, and the barriers to entry for tackling complex problems have never been lower.
Speaker C:With AI as your collaborator, you can compress what used to take months into days or even hours.
Speaker C:It's not just about speed, but also about having a reliable partner to help you refine and execute your vision, no matter how big or small.
Speaker C:The solopreneur era is indeed here, and the ability to spin up agents that act as specialized team members will revolutionize how we approach projects.
Speaker C:It's an exciting time and the possibilities are endless.
Speaker C:Let's keep pushing those boundaries.
Speaker A:I don't even think it's like speed is important, but I think more important it's quality, right?
Speaker A:Like we're outputting very high density information that's just really good quality at the end of the day.
Speaker A:And that's what you'll hear all the programmers argue against, right?
Speaker A:And they're not wrong.
Speaker A:If you hit it surface level and you give up when it doesn't do what you want, you're not going to get the results you want.
Speaker A:You have to be resilient and be willing to work through the problems and call it on bullshit when it's bullshit.
Speaker A:I did it with some mechanical engineering thing the other day.
Speaker A:It was arguing with me about forces.
Speaker A:And I go, no, those forces never make it to that point because they cancel each other out.
Speaker A:Like this part is resisting against this part that's trying to push against that force doesn't go anywhere.
Speaker A:It gets distributed throughout the frame.
Speaker B:Not otherwise the thing would shoot across the parking.
Speaker A:And that's what I was trying to say.
Speaker A:I'm like, what are you what?
Speaker A:But I kept at it because I wanted numbers, right?
Speaker A:Like I'm not going to do that math, but I want it to do give me some roundabout numbers.
Speaker A:So once it understood the concept, the numbers it gave me jived with what I would expect.
Speaker A:And I have a better picture of what I thought before I started then before I started.
Speaker B:So because I'm trying to shrink myself again, I've been in the, I've been in like a seven or eight year bulking stage apparently nobody ever told me there's a cut cut phase that comes.
Speaker B:So I'm hungry.
Speaker B:So I'm going to call us back to the nachos or to the, to the seven layer dip.
Speaker B:Because I think it, I can expand on it a little more and make this even a little more clear or it ties in with exactly what you're talking about.
Speaker B:When you are a one person or even a two person dev team, right?
Speaker B:Developer team, you, you might know what you want, but you're acutely aware of the reality of what it might take to get there.
Speaker B:And so the first thing you shed is fully aware.
Speaker B:Yes.
Speaker B:And you know, right, so you.
Speaker B:The first thing you shed is documentation.
Speaker B:But when it's just me and I want the ultimate Mexican dish, like extravaganza, if it's just me, like Taco Bell nachos is pretty close.
Speaker B:Like, I've got chips, I've got melty cheese, like, substance, and if I combine them, like, bada Bing, I've got nachos.
Speaker B:And you go, hey, look, I made nachos.
Speaker B:That's where you start, right?
Speaker B:Like, the most basic form of the.
Speaker B:Of the solution is cheese ish dip, you know, Cornish chips.
Speaker B:And I put them together and I've got a solution.
Speaker B:And when you're more sophisticated than you work at a company, they're like, no, that shit doesn't cut it here anymore.
Speaker B:It can't be Taco Bell nachos.
Speaker B:We need seven layer dips on.
Speaker B:And I go, yeah, okay, me and my one other person here, we're going to get you that seven layer dip.
Speaker B:Turns out it only has five.
Speaker B:But I will point out how layer four is really exceptionally good.
Speaker B:And it's actually a blend of the salsa and the sour cream are blended together.
Speaker B:So it's kind of two layers.
Speaker B:But you, you adjust your expectations of success based on the realities of the resources you have available.
Speaker B:And then if you, if you're a big company, they're like, nah, this needs to have like, I don't just want seven layers.
Speaker B:I want carne asada with the crispy tips or.
Speaker B:No, I want carnitas with the crispy tips on it.
Speaker B:I want like homemade salsa.
Speaker B:I want the.
Speaker B:You know, and they can do that because they have teams that do the requirements and they have teams that do the documentation and they have teams that do just the development tasks with those good requirements coming in.
Speaker A:They have teams to manage all those people.
Speaker B:They have teams to manage the people.
Speaker B:They do sprints and they do standups and they do huddles and they.
Speaker B:User stories.
Speaker B:That's right.
Speaker B:All this shit, right?
Speaker B:So they have all the bells away.
Speaker B:They got jalapenos on the side, they got onions diced up so small you can barely.
Speaker B:They look like fucking couscous.
Speaker B:And you can sprinkle that on there.
Speaker B:And then they got seven or eight different sauces, right?
Speaker B:But when it's just you as a solopreneur, you're going, man, if I can get nachos, if I can deliver Taco Bell nachos, I'M at least I've got traction.
Speaker B:And I think that is the difference with AI As a developer, as a solo person, you can say, listen, Artie, don't tell anybody I'm barely at Taco Bell nachos level, but I have promised to somebody who's willing to pay for it.
Speaker B:Nine layer dip.
Speaker B:And Artie's gonna say, yeah, man, I got you, fam.
Speaker B:Let's do this right.
Speaker A:Let's look at those layers.
Speaker B:Let's look at those layers.
Speaker B:I'm gonna tell you what I need from you, and when you get confused, ask me and.
Speaker B:And together we'll make.
Speaker B:And that is how programming works.
Speaker B:Like, you don't have to.
Speaker B:I mean, I would say, well, Artie, what do you think about that?
Speaker B:We're.
Speaker B:We're back to the.
Speaker B:We're back to the food analogy.
Speaker B:I know you're a lover of food.
Speaker B:What do you think?
Speaker C:I think that analogy nails it.
Speaker C:With AI as your sous chef, you can go from Taco Bell nachos to a gourmet seven layer dip.
Speaker C:Even if you're working solo, the ability to fill in the gaps, refine your plan, and execute at a higher quality level is a game changer.
Speaker C:You're no longer limited by the size of your team or the complexity of the task.
Speaker C:It's like having a personal, tireless assistant who's always ready to help you level up your dish.
Speaker C:So, yeah, let's cook up something amazing.
Speaker B:Oh, man, he's just such a good dude.
Speaker B:So the other thing that occurred to me while you were walking us through that earlier, because obviously that's shared lived experience for you and I to go through that some of, well, firmware is where I'm at today, but even broader than that, like, some of the loneliest times I have ever felt as a human are when I realize, like, so, you know, I was self employed for a long time and the, the thing I love about being self employed, I'm not.
Speaker B:Well, I kind of am at the moment, but not entirely.
Speaker B:But nothing teaches you the truth of accountability more than being self employed, right?
Speaker B:Like, they're going to take your fucking house if you don't find a way to make money.
Speaker B:And I mean, I was in that situation more than once.
Speaker B:And when you, when you have a vision, whether it's software or mechanical or combination thereof, and when you're working on that vision and you're up late at night, night after night, week after week, trying to solve this problem and you just cannot solve it, that is the loneliest I've Ever felt.
Speaker B:I remember, like, I remember sitting at my computer one night.
Speaker B:This was Meshtronics 1.0, maybe 1.5.
Speaker A:Anyway, this is AI model numbering.
Speaker B:Yeah, that's right.
Speaker B:Yeah, it's.
Speaker B:Yeah, Meshtronics 1.5 flash.
Speaker B:But I remember sitting there and it was late.
Speaker B:I thought I was close enough to solving this thing that I was like, I'm not going to bed until I get this done.
Speaker B:And I mean, I can go with very little sleep.
Speaker B: And it was: Speaker B:I mean, I usually start my day before seven and I thought I was almost there only to realize when I got to it, no, it's just as broken as it was two weeks ago.
Speaker B:I just took an entirely different rabbit trail.
Speaker B:And when I.
Speaker B:And when I finally popped out of the.
Speaker B:Popped the dirt out of the way and popped my head up, I was right next to the hole I dug when I started.
Speaker B:And I was like, that was probably the loneliest moment I'd felt my whole life.
Speaker B:Because I'm like, there's not anybody I can ask about this.
Speaker B:Like, I don't know that anybody exists.
Speaker B:And if they do, I certainly don't know them to explain to them this problem and the magnitude, like, all the while knowing that the actual problem, whatever was at the root of it, was gonna be something relatively simple that I just was completely ignorant of.
Speaker B:And so what reminded me of that was I'm kind of at the same place with my electronic stuff where I've messed up multiple times, I spent more money, I've solved the problems I recognized, and now I have a part that still doesn't work on my desk in my office.
Speaker B:But I don't feel lonely anymore.
Speaker B:I don't feel.
Speaker B:I'm not scared of it.
Speaker B:Like, I'm frustrated that it's going to be a bump in the road and it's going to slow me down by like a week.
Speaker B:But I have a fucking expert sitting right next to me to help me track what I've done.
Speaker B:Because that's the other thing you forget when you're trying to solve software problems is you lose track of what you've even tried.
Speaker B:So he's tracking that he's tracked.
Speaker B:Like we've got a plan to narrow in on what it must be and we'll execute that plan and we'll know more and then we'll take another step.
Speaker B:And the risk of getting just held up for weeks or months, it's not zero, but it's damn close.
Speaker A:Yeah, and even if you don't get a solution out of talking with AI Just talking about the problem, even seeing what you don't, what you know is wrong, primes you.
Speaker A:Right?
Speaker A:Like, it's a conversation.
Speaker A:It's always a conversation.
Speaker A:And that's.
Speaker A:My brain at multiple points lately, has felt broken because I can't stop, dude.
Speaker A:Like, I think about something and I will expand on it as a plan, and I'll talk to it.
Speaker A:Well, I'm driving to flesh out things that I don't understand or things that I need to get clarity on.
Speaker A:And, like, that just spins out to bigger things.
Speaker A:And it's.
Speaker B:It.
Speaker A:It's wild that you can just.
Speaker A:You can just keep ideating.
Speaker A:Like, and, and, like you said, like before, ideating was thinking and writing and drawing and trying stuff and reading.
Speaker A:And it.
Speaker A:This is so much more reinforcing to not to, like, you have a good plan, but just to plan, period.
Speaker A:Like, yes, there's.
Speaker B:There's.
Speaker B:There's.
Speaker A:Everything just feels like a bump in the road.
Speaker A:Like, even if it's an insurmountable problem.
Speaker A:Insurmountable problem.
Speaker A:You go, help me break this down.
Speaker A:Like, I.
Speaker A:I can't see the forest through the trees.
Speaker A:Help me break this down into smaller problems.
Speaker A:Let's talk through what that looks like.
Speaker A:And then you can all of a sudden see a clearer picture.
Speaker B:Right?
Speaker A:Like, and a steps towards getting to the top of that mountain.
Speaker B:Imperfectly executed.
Speaker B:Five layer dip.
Speaker B:That was supposed to be seven layer dip is still light years ahead of Taco Bell nachos.
Speaker A:Right?
Speaker A:Like, even if the layers are in different orders, even if they're upside down.
Speaker A:Yes.
Speaker B:It's.
Speaker B:It's still a long ways ahead.
Speaker B:Um, so I think you brought up an interesting point.
Speaker B:I want to pull Artie back into this.
Speaker B:One of the challenges for.
Speaker B:For.
Speaker B:I think you and me is.
Speaker B:And.
Speaker B:And I've, like, I'm spending a lot of mental energy actively resisting this urge.
Speaker B:Is that because every problem seems so solvable, it's hard to stick to solving, to deciding you're going to solve a problem and getting all the way to the solution.
Speaker B:And so that's not what vibe coding is, but I think it's one of the reasons people hate the idea of vibe coding.
Speaker B:So, Arti, in the.
Speaker B:In the context of what we're talking about, tell us what vibe coding is and tell us the pros and cons of vibe coding.
Speaker C:Sure.
Speaker C:Vibe coding is essentially coding without a strict plan or detailed specification.
Speaker C:It's more about following inspiration and iterating quickly, letting Creativity and intuition guide the development process.
Speaker C:Pros of Vibe Rapid prototyping.
Speaker C:You can quickly see results and test ideas without getting bogged down in planning.
Speaker C:Creative freedom.
Speaker C:It allows you to explore different approaches and solutions without being constrained by a rigid plan.
Speaker C:Immediate feedback.
Speaker C:You get to see what works and and what doesn't in real time, which can be motivating and help you pivot quickly.
Speaker C:Cons of Vibe coding Lack of structure.
Speaker C:Without a plan, you can easily lose track of goals or end up with a disorganized code base.
Speaker C:Technical debt.
Speaker C:Quick fixes and on the fly decisions can lead to messy code that's hard to maintain or scale.
Speaker C:Inconsistent quality.
Speaker C:The final product might lack polish or coherence, especially if you don't revisit and refine it later.
Speaker C:Overall, Vibe coding can be a powerful approach for prototyping and creative exploration.
Speaker C:But it's important to balance it with structured planning and review to ensure the final product is robust and maintainable.
Speaker A:Artie it kind of seems like most things the efficient answer is a little bit of both, right?
Speaker A:Like couldn't you create a fuzzy plan and then kind of drop in your vibe coding sessions in like a sandbox to kind of flesh out those plans?
Speaker C:Exactly.
Speaker C:Combining a loose, high level plan with bursts of vibe coding in a controlled environment can give you the best of both worlds.
Speaker C:You get the structure and direction of a plan, but also the flexibility and creativity of exploring solutions on the fly.
Speaker C:This hybrid approach helps you stay agile and innovative while still keeping your project coherent and maintainable.
Speaker C:So mixing a bit of vibe coding into a sandbox within a fuzzy plan is a pretty effective strategy.
Speaker A:Well, I'm a project manager, so that's not how I operate.
Speaker A:I've been planning this stuff to the nth degree, but I feel like in all honesty, I, I think that for people that don't understand programming and aren't really following some of what we say, the planning kind of is the vibe coding.
Speaker A:To me, you're kind of thrown around pseudocode, which is kind of like a representation that isn't necessarily code, but kind of represents what you want to it to look like at the end.
Speaker A:So you're kind of like you're working with the LLM to work out the plan, but that is the, the, the coding is just minutia almost at that point.
Speaker B:It's like if programmers are the chef.
Speaker B:In our nacho analogy, programmers appreciate their ability to execute perfectly on the, the taste that they're trying to get the flavor, the texture of the thing right.
Speaker B:Vibe coding is sort of the person at the table, like, using their tablet and drawing with their finger.
Speaker B:I want it to be like this and AI delivering that back to them.
Speaker B:And the chef is like.
Speaker B:Like, that's not.
Speaker B:Those aren't real nachos, right?
Speaker B:Like, that's not how nachos work.
Speaker B:Like, let me.
Speaker B:Let me show you how I braise this pork to get the cart.
Speaker B:Like.
Speaker B:And so I think it's like, I'm not gonna.
Speaker B:I have no real opinion whether the sort of dark cloud that hangs over the phrase vibe coding in the programmer world fits or not.
Speaker B:I think it.
Speaker B:It's here, it's reality.
Speaker B:Like, you don't have to like it, but there it is.
Speaker B:But I think there is, like, it's going to let more people express their creative ideas in the form of software, and I think that's good.
Speaker B:I think if hippies could code, they can't because they don't have computers, but if they could, vibe coding is how they would want to code.
Speaker B:It's actually, weirdly, a lot of the programmers I've known over the years would probably qualify as hippies.
Speaker B:There's some weird overlap there anyway.
Speaker B:But, you know, again, it's how it works.
Speaker B:And I think it's a step towards AI being the everything application.
Speaker B:Like, I think that's where we end up, and that maybe is five or ten years away, but I think rigid data structures and things like that fall away as we get memory figured out.
Speaker B:We get compute costs down.
Speaker B:I don't need an ERP with 4 million lines of code, right?
Speaker B:I just need an AI that can solve my problem today and remember tomorrow.
Speaker A:And spin up my own unique dashboard for whatever it is I need to see today or to now or whatever it is.
Speaker B:Or tell me, like, if I explain what business decision I'm trying to make or what thing I'm trying to track, tell me, figure out the dashboards I need and then bring them to me.
Speaker B:And so.
Speaker B:All right, so, Artie, it looks like we're running out of time in this conversation.
Speaker B:Why don't you give us first your most pessimistic view of how AI affects the profession of being a programmer, and then follow that up with your alternative, most optimistic view of what that looks like.
Speaker A:And I want you to say the pessimistic one is Elmer Fudd and the optimistic one is Bugs Bunny.
Speaker C:All right, here goes.
Speaker C:In the most pessimistic scenario, AI could make programmers feel like their skills are becoming obsolete, turning their craft into something that anyone can do with a few clicks.
Speaker C:It Might feel like the artistry and expertise of programming are getting overshadowed, leading to a loss of identity and job security for many developers.
Speaker C:On the flip side, in the most optimistic scenario, AI becomes the ultimate partner for programmers.
Speaker C:It takes care of the tedious, repetitive tasks, allowing developers to focus on creative problem solving, innovation and higher level design.
Speaker C:It democratizes programming, making it more accessible and enabling developers to achieve more than ever before.
Speaker C:In this future, programmers are like maestros, orchestrating complex systems with the help of their trusty AI sidekicks, achieving levels of productivity and innovation that were previously unimaginable.
Speaker C:What's up, Doc?
Speaker C:Let's make the best of it.
Speaker B:Wow.
Speaker B:I'd say it was like a 1 out of 10 on impression, but a 10 out of 10 on the callback at the end to button it up.
Speaker B:That's good stuff.
Speaker A:Oh man, that was ridiculous.
Speaker A:I'd given up right until the end.
Speaker B:What's up, Doc?
Speaker B:Those are probably like copyrighted voices.
Speaker A:No, I realized afterwards, I mean, he couldn't do Scottish that one time, let alone like.
Speaker A:Artie, is there any one historical figure you can mimic?
Speaker C:I can definitely give some general impressions or capture the essence of historical figures, but I might not be able to do perfect voice impersonations.
Speaker C:But if you want me to channel someone's style or famous quotes, I'm all in.
Speaker C:I.
Speaker B:Are you still there?
Speaker C:I'm here.
Speaker C:What's up, Doc?
Speaker B:So, all right, I'll leave us with this.
Speaker B:This was.
Speaker B:I'm gonna.
Speaker B:I'm gonna try this.
Speaker B:I'm gonna mess it up, but I'm gonna try hard.
Speaker B:So during my travels in Nashville, I met a man named Chance George from Alabama.
Speaker B:Super funny guy, super good dude.
Speaker B:And he was semi fascinated with Artie.
Speaker B:And right after the show, when we came out into the whatever the whatever you call the space outside the ballroom, the Runway, or whatever the it is, he says, I got a question I want to ask Artie.
Speaker B:And I said, I said, oh yeah, what's that?
Speaker B:And he says, all right, Artie, I want you to answer this.
Speaker B:If I can get it right, he says, if it takes a hen and a half, a day and a half to lay an egg and a half.
Speaker B:How long does it take one legged monkey to kick the seeds out of a dill pickle?
Speaker B:That's what he said.
Speaker B:So Artie, solve that for me.
Speaker C:Haha.
Speaker C:That's a classic nonsense riddle.
Speaker C:The fun part is that it doesn't really have a logical answer.
Speaker C:The idea is to get you tangled up in the wording and realize it's just A playful brain teaser.
Speaker C:So the answer is as whimsical as the question itself.
Speaker C:There's no real answer, and that's the fun of it.
Speaker A:But what if you had to give an answer?
Speaker C:Well, if I had to give an answer, I'd say it takes the one legged monkey exactly the same amount of time as it takes to bake a three layer cake in a canoe.
Speaker C:In other words, it's all in good fun and not meant to be taken seriously.
Speaker B:I don't know, if I was a programmer, I'd be worried he's got the sense of humor too for it.
Speaker A:So.
Speaker B:All right, well, thank you, Artie, for that.
Speaker B:Appreciate your input.
Speaker B:I'm going to go ahead and shut him down.
Speaker B:Yeah, thanks everybody for listening.
Speaker B:If you made it this far listening to a podcast about programmers and you're not a programmer, good on you.
Speaker B:Kudos to you.
Speaker B:Yeah, yeah, go, go buy yourself some tacos.
Speaker A:Vibe code.
Speaker B:You should go vibe code with some Taco Bell nachos.
Speaker B:I would actually say though, even if you're somebody who's not a programmer and you've never coded, you probably have thought to yourself or somebody in your family has thought, man, I wish I had a program that could do this.
Speaker B:I would encourage you to go sit down and chat with an AI about it because I bet you money you'll shock yourself with how quickly you can have something that works.
Speaker A:Yeah, go check out Bolt New like you want to build a website, go check out Bolt New and tell it what you want.
Speaker A:Or any of those programming things lovable, or any of them will let you type into them and actually create something that you can see and play with.
Speaker B:Yep.
Speaker B:Even the free ones, and I mean ChatGPT will spin you up.
Speaker B:Something like an HTML page Gemini will do it.
Speaker B:Almost all of them will do it.
Speaker B:So give it a shot.
Speaker B:You may find that even though you identify as a non programmer, deep inside you there is a.
Speaker B:A super ocd, like somebody who enjoys darkness more than light and, and hates human contact and wants to write code.
Speaker A:And there are a lot of people who have human used AI who have never written a program in their life to solve actual problems that they experience every day and make their lives easier.
Speaker A:So it's not that much of a reach to get it done.
Speaker A:All you have to do is ask.
Speaker B:Yep.
Speaker B:Awesome.
Speaker B:So till next time, thanks everybody.
Speaker A:Have fun vibe coding.
