← Resource library

3 Hot Takes on AI, Vibe Coding & the Future of Software Development

Watch on YouTube

About this video

AI is changing the way we build software, run businesses, and work with technology. But are we getting too caught up in the hype?

In this video, I'm sharing 3 hot takes on the current state of AI and software development, based on what I'm seeing with real clients, fellow freelancers, and agency owners.

As AI tools become more powerful, we're seeing new possibilities for automation, software development, and building custom tools. But that doesn't necessarily mean developers are going away or that every AI-generated app is worth paying for.

Here are my 3 hot takes:

1. Everything is becoming compatible with everything.
AI is making it easier than ever to connect systems, automate workflows, and build things that previously required complex integrations. From AI-powered game modding to everyday business automation, the possibilities are expanding quickly.

2. We're still going to need software developers.
AI can write code, but domain expertise still matters. Experienced developers know how to guide AI, solve problems, and build reliable solutions. And sometimes, using an established platform like Webflow is still better than maintaining a custom-built system.

3. Ideas don't matter. Execution does.
When almost anyone can build an app with AI, simply having an idea isn't enough. The real value comes from how well you execute, distribute, and build a community around what you create.

We'll also discuss the rise of vibe coding, the hidden costs of custom software, intellectual property in the AI era, and why building an audience might be one of the best investments you can make.

AI is creating incredible opportunities, but knowing what to build, when to use existing tools, and where your expertise adds value is becoming more important than ever.

What do you think? Are developers becoming more valuable in the AI era, or will AI eventually replace the need for them?

Chapters

  • 00:00 Introduction
  • 00:44 Hot Take #1: Everything Is Compatible
  • 03:07 Hot Take #2: We Still Need Software Developers
  • 06:04 Hot Take #3: Ideas Don't Matter, Execution Does
  • 09:03 Final Thoughts

Full transcript

2,318 words

Copy transcript

Today, I want to share some recent thoughts I've had about the broader state of AI and software development. I work with clients every day, and I talk to my peers and other freelancers pretty regularly. I know that my peers can be in a bubble. So, I'm trying to bring a little bit of a balanced take between what I see in the real world with clients, what I see online, what I talk about with other freelancers and agency owners. The thing about me and my peers is that we are often early adopters.

We always want the new technology, we're always using new workflows and new tools to try to make things as fast as good as possible for our agencies, but that tends to put us in a little bit of a bubble where we are not experiencing the real world the same as everybody else. And it's easy to think that we're be even behind because things are moving so fast right now. I've got a couple of thoughts, so here are my three hot takes.

Take number one is that everything is compatible with everything. Almost anything in the digital world can be connected or transformed, automated, recycled, mismatched together, and it can be created by AI from one system to almost any other system, no matter where it lives, honestly, whether it lives on your computer or whether it lives online. Even if it doesn't have an API or any other way to get to it, if you can get to it on your computer, AI can even use your computer and your mouse to extract data and to automate it.

It can even be running in the cloud overnight while you're asleep. Now, it's true that this process isn't always fast or accurate or even super cheap if you're using an expensive coding model, but it's not all those things just yet. I still think that we have quite a ways to go where this tool is going to be better and better and better, and I know some people don't really think it can continue at the speed that it is, whether it's like token economics or just speed of development in general.

My opinion is that the market is still pushing people to operate really fast, and there's still good competition between some of these big companies, and people have an appetite for this stuff. So, I think as long as there's people willing to pay, that there'll be people who are producing this at a really fast rate. A concrete example of this is what's happening in the Halo world and Xbox as a company. Recently, modders have been going kind of nuts with AI and Halo files and the like decompiled versions of the game and they're putting Halo and Halo cars and all that into Modern Warfare 2.

There's another project that has Halo 3 running in your browser with up to 128 people, which is pretty awesome. Now, to their credit, instead of coming down with a big legal banhammer and deleting this project, Xbox and their leadership has set up a new subsection of their studio called Xbox Labs. And if you read between the lines in their announcement, it seems like they're pretty aware of this community that's using AI and mods and all these new technologies to accelerate and to invest in their game.

It's really cool that they're coming alongside this community to try to help foster and capture this enthusiasm instead of just coming in and saying, "You can't do this. You can't use Halo IP. You can't use our Halo characters like this." I think it's a pretty bold move because this is moving really quickly and it would be easy to just shut it down and say, "No, you're not allowed to do that." I think it's a bit of a leap of faith for Xbox to say, "Well, let's see where this goes." So, really, almost everything in the digital world is compatible or will be compatible with almost everything else.

So, that's a more extreme example, but really, when it comes to your business, you can say, "Look at my email, find the PDF of all my kids' school events and put it on my Google calendar." Right now, that's really kind of a trivial task for AI and even a couple months ago, that would be pretty impressive, but now the models can do that almost no problem. My second hot take is that people are still going to need software developers.

And I think it really just comes down to domain expertise. There's a great video by Jamie Brindle does a skit where he pretends to be the client and then he's the expert freelancer. And the client says, "Hey, I'm sorry, we're going to pass on you. We decided to go do it ourselves with AI." And so, the freelancer says, "Okay, great, no problem. How about this? You do a version with AI and I'll do a version with AI and we'll see which one's better." And the client goes back and says, "Oh, well, that's not fair.

Yours is obviously going to be better cuz you're the professional." Oh, wait a second. It's a funny skit, but I think it illustrates the point really well that the expert is going to get better results in the end. And I think a lot of that has to do with just the technical language and the more experience that these software developers have. Even just knowing how to tell the model what is wrong. You can waste a ton of time just going in circles saying, "That's not right.

That's not right." where the expert could say, "This is failing in this specific way. Here's what you need to do instead." And that just is going to create a much better result from the model. I think a lot of the models are just going to have that problem forever. And I really do think that that's almost always going to be the case unless you're doing something like lovable or bolt where a lot of the technical decisions are already made behind the scenes.

And then you are pretty much siloed into what results you're going to get. I think experts also have more powerful tools. Whether it's software or in my case, I have video cameras and things like that. Experts are going to have just a better result in the end than an amateur or somebody who's using AI for the first time. One analogy might be just because we all have 3D printers available at our fingertips, doesn't mean we're all making our own toys.

And speaking of that, a funny example from my own life, I'm trying to print a little battleship game for my kids and it's taken me weeks because there's all these little pieces that have to be printed individually in different colors. And sometimes the prints fail or they don't fit right, even though I think I know what I'm doing, doesn't work exactly how I want the first time. That just takes a ton of time and I have a full-time job.

Really, it's just not practical for me to be printing games for my kids all the time. And I think for a lot of customers, it's not practical for them to be vibe coding and building all their own tools. This tweet here says, "Just met a company that vibe coded an entire CRM to avoid paying for HubSpot or Salesforce over the last year. It's now become a burden to maintain and missing key functionality and third-party interoperability as they scale.

They're now migrating to HubSpot." And he says, "App layer is fine." Which is sort of his way of saying some things are just better to outsource and it's not going to be a problem for software developers. And I think I agree. I think there's plenty of things that are not practical to actually build yourself. Now, as a side note, this is why I have a lot of confidence that Webflow is still the right path for a lot of customers.

My personal opinion is that there's going to be a lot of paper cuts and hidden costs that pile up over time with all these custom systems. Even before vibe coding, I was seeing this with clients that inherited a custom product or were on a custom stack and they didn't like it. They wanted to be on Webflow because it was a lot easier to use. Now, I think Webflow, again, has solved a lot of these problems that marketers face. I really I we're going to see the pendulum swing back.

Especially if we get another round of marketers that inherit a vibe coded site. I think they're going to look at this and say, "I hate this. Get me out of this. Get me back to Webflow or get me back to some tool that I know how to use. This is too too much custom stuff for me to handle." So, I think all those little papercuts are really going to pile up. And I think developers are really still going to have a huge place in this world.

Here's my last take. Ideas don't matter, only execution. I say this meaning no shade at all to my Webflow friends who have shown me their vibe coded tools, but I personally am just not willing to pay for a vibe coded app for my business. I don't want to be paying for a vibe coded app that I could do myself in just a couple of hours. The problem really is that these tools are so easy to create that there's really no barrier to building it myself.

You may have given me the idea, but I can just make that same tool in Claude in 2 hours. My version will be more to my liking and it'll be a better fit for what I'm actually trying to do in that case, and it'll have my colors and my design and and I'll I'll be able to own the code and make the changes that I like. And unfortunately, we now live in a world where people can reverse engineer your product and even your game in the case of Halo.

Everybody can reverse engineer everything else, and theoretically they can just reverse engineer your entire product, steal your software, and build it for themselves. Either they wouldn't be a customer cuz they're just cloning your idea and building it for themselves. Or in the worst case, they're going to clone your idea and try to steal your customers. Unfortunately, I think this is kind of the new world that we've entered.

People are behaving like intellectual property just doesn't exist. It's true that copyright protections right now are extremely murky, and it really doesn't help that all these models have been trained on every bit of copyrighted content in human history. From stock photography to books to blog posts to YouTube videos, all that stuff has been sucked into these models that we're now using to remix and create new things ourselves.

So, it's really murky, and I understand that it's hard to nail down who actually owns this idea or this piece of content. There's a video my brother showed me years ago, but it's really stuck in my brain, and I know it's a bit controversial, but I'll put it up on stage here. The premise of this video essay is that the author thinks that there should be no copyright protection laws at all and everyone should just be paid via like a Patreon or a Kickstarter model.

And beneath all that, the baseline idea is that the only things that are valuable either the production of the idea or the distribution of the idea. I actually think that there's a lot of truth in that, especially in this day and age. The example in the video was if there were no copyright protections or intellectual property laws, anybody could make a Star Wars story, but fans would be most willing to pay for a Star Wars story from one of their favorite creators or somebody who's done another a great movie in the past.

They would want to see that person's vision of Star Wars without having to go through Disney or Lucas Film or any of those things. Theoretically, the argument was that the money would flow towards whoever was actually the best producer of those ideas. I know it's a bit of a rabbit hole, but here's my point. I think you should definitely be building internal tools and workflows to make your own execution better, but I really wouldn't recommend trying to build a micro SaaS product that you're trying to get people to pay for in this day and age.

I also think it's really important to be building an audience and a community. I think that's a great return on investment because people want to hear your ideas and want to be connected with you as a personal brand. And that's super valuable. That's one of the reasons I'm doing this these videos in general. It's my opinion that that's what's going to set you apart a lot more than just what features your app can have because those kinds of things can be stolen or copied or replicated really easily.

I think you should be giving away your ideas as much as possible because I think it's the execution and the production of those ideas that's going to set you apart. Those are my three observations about the state of the industry right now. Thank you for watching and for listening to my ideas and I'll see you next time.