[00:00] I'm done with the AI hype. Now, I know this may sound like an old man yells at clouds video. I mean, I am old and AI agents live in the clouds, so there's that, but it's not really a video like that.
[00:15] You probably noticed that I don't cover AI assistance development all that much on this channel. It's actually on purpose. It's not that I hate AI and agents or something. I don't hate it at all, actually.
[00:27] I use it all the time myself. But I do think it's important that we are honest about what AI can and what it can't do. And what I'm mainly fed up with is the limited view that some people seem to have
[00:41] about what software development actually is. I think in some places AI helps us, but also it's not going to fully replace what software development is. But I'll talk more about that in a minute. Now, I also have some announcements about the channel.
[00:55] Some things are going to change a bit. and want to share that with you. Now, what is software development? There's a lot of people posting online about how they create complete teams of agents with different roles, engineers, staffers, managers,
[01:09] agents that engage in office politics, annoying colleague agents that keep bothering you. And with that team, they then launch complete applications within an hour, ship features within 10 minutes.
[01:21] Because that's what software development is all about, right? Shipping lots of stuff really fast. And whoever skips the most features the fastest wins the game. But is that really true, though? Is that really what matters?
[01:35] I mean, if you're a startup company, you want to develop an MVP and share it with your users to learn very quickly where your idea is valid. I get it, that's important. The faster you know that, the better.
[01:47] But when you think about it, that's actually just a small slice of what most software development actually entails. Most software runs inside big companies that's doing all sorts of crucial work that you can't
[02:00] just change on the whim. If you work, let's say, at a big bank or an insurance company, you don want to shift things fast That very risky It reckless All sorts of thing will run You can expose secrets lots of customer data that in danger accidentally delete important information
[02:18] seriously breaking a crucial part of your backend that's going to cost you millions. And you end up with a bunch of lawsuits. Big companies don't care about shipping fast. They care about minimizing risk.
[02:31] And that's the first issue I have. You can't blindly apply what is important to a solo developer for their side project, or a small team or an AI startup to a bigger established company.
[02:43] The needs are going to be very different. Of course, bigger companies also want to use AI as part of their software development, and they should. But a lot of other things there need to happen as well. And lots of different people are going to be involved, from finance to risk management, marketing, sales, you name it.
[02:59] And that points to a second issue that I have for these types of posts. It presumes that software development equals writing code. Of course we write code.
[03:12] But honestly, it's a small part of what we actually do. We look at design and architecture. We organize things. We talk with stakeholders to figure out what we want. We write tests. We spend time coming up with creative solutions to problems.
[03:26] We think about what it means to design a system in a certain way. Can we design things so they are secure by default, that they naturally promote the way we want users to interact with our systems? How can we make things simpler, more pleasant to use?
[03:39] How can we make sure that what we've built today doesn't have to be done again tomorrow or at the next sprint? And honestly, I view writing the actual code as an afterthought. If you thought things through beforehand,
[03:54] writing the code is not the hard part. I'm not saying we should go back to the waterfall model, we definitely shouldn't, but setting up things properly, rethinking things through, simplifying systems, coming up with creative generic solutions that stand the
[04:07] test of time, that is the hard part. It's like saying that you are a carpenter because you have an impact driver. You know it's nice to have a machine that drives screws into wood but can you frame a roof Can you build a wall Driving the screws into the woods is the easy part Now you might say well you know that where AI is now In a few
[04:27] years, it's going to do all those things as well. But here's the thing. If you don't have any business experience, you might think that running a business is about delegating tasks. If that's the case, then yes, if AI can do all the tasks, well, we don't need people
[04:42] anymore. But that's actually not true. Running a business is not about delegating tasks. It's about delegating responsibilities. As a social developer, you're not just doing tasks. You are responsible
[04:56] for things. If you use AI to ship a feature that seriously damages your company, AI is not responsible for that. You are. And your boss is going to hold you responsible and you're going to
[05:08] get fired. That's not going to change. Somebody needs to be there to assume responsibility, decide what needs to be done, and check that it's done correctly. So when you think about it that way, what are actually the skills that you need to be a successful software
[05:23] developer? Well, in short, you need to be someone that can take responsibility for things. And you can only do that if you know the fundamentals. To build a house, you need to know the fundamentals
[05:36] of carpentry. Being able to drive a screw into the wood is not the hardest part. And similarly for us to build software, you need to know the fundamentals of software design and development. Because writing code is not the hardest part. And that's also the reason
[05:53] that I don't spend too much time on my channel talking about AI agents. Because that's mainly related to writing code and it's not that interesting to me. I care about fundamental design principles like cohesion, coupling, abstractions and how you translate those
[06:09] things practically to the real world. And this is also what I will keep doing, what I will keep teaching on the channel. I don't want to make this a fluff or AI hype channel that you know, oh we have a GPT-5 point whatever that's going to completely destroy Gemini or whatever other
[06:26] AI model we have this month I don care about those things I focus on the fundamental things that I think you need to know to be a great software developer That being said from now on I might take a small break
[06:40] from time to time. I feel like I've been grinding really hard over the past years, always publishing a video every week. It's been very intense actually. Don't get me wrong, I love doing it and I
[06:53] I really appreciate you being here and watching the content and sharing your thoughts. It's amazing and I really feel lucky that I'm able to do this for a living. So I'm definitely not quitting.
[07:05] There's a lot still for me to cover on the channel. But, you know, sometimes there might be a week without a video. Well, I'll take some time to reset, come over from where it is, and try to make the videos even better for you.
[07:18] Finally, if you are enjoying these videos, you want to dive in deeper, into this topic, I am launching a new program this year called Software Design Mastery. This is actually how I think software design should be taught. It teaches you the fundamentals
[07:33] of design through videos, exercises, quizzes. It takes you step by step through working on improving simple code to properly designing complete systems, but also dealing with the
[07:46] trade-offs of design, because there are many trade-offs. There's going to be a community, There will be lots of extra content added to the program. It is going to launch very soon. The doors will be open for only a few days, and then it's not going to be available for several months.
[08:00] So if you're interested, make sure you join the waiting so you don't miss out. The link is in the description. So AI, yes, use it to be more productive and let it write code for you. But don't forget, you are responsible.
[08:13] You need to know about the fundamentals, and that's what really matters. But I'd love to hear what you think. Do you care about the fundamentals of design? Do you think that AI agents are going to be able to do those kind of tasks?
[08:27] What are your thoughts about how everything is going to change in a couple of months or years? Things are moving so fast nowadays. It's lovely to hear from you. If you want to learn more about software design, check out this playlist right here.
[08:41] Thanks for watching and see you next time.