Python Has Serious Problems, But I Use It Anyway

Confession time. I used to not like Python at all and now I run a pretty big Python focused YouTube channel. So either I completely changed my mind or perhaps something more interesting happened. Today I'll tell you why I didn't like Python and why most of the things that I didn't like about it are actually still true and why I use it anyway. And actually there's a bigger lesson here about how you think as a developer. So, no code examples today. It's just me sharing my thoughts about Python and what I think about it. By the way, I'm currently building something new called software design mastery. This goes much deeper into software design and architecture than I can do on YouTube. If you want to become a more confident senior level developer, make better design decisions in your code independent of the programming language, join the free weight list at iron. co/mastery. The link is in the video description.

Now before I want to talk about Python, let's put this into context. Actually when I learned programming, I didn't start with Python at all. I actually started with basic on a Commodore 64 uh saving programs on cassette tapes. Yes, I'm that old. And then I went to Turop Pascal and then in high school I learned C++. Uh at university I learned modular 2 and then Java. When I became a teaching assistant, I built some simple web systems in PHP, HTML, JavaScript and during my PhD studies, I focused heavily on C and C++ for well computer graphics which was the topic that was that I was working on at the time and it had Python as scripting layers. That was actually the first time I came into touch with Python. After finishing my PC, I started working as a teacher and there I taught C. So yet another new language for me teaching game programming and later on in my businesses startup companies and even iron codes I used things like uh react and typescript for front- end work python for tooling and automation and javascript nodejs for the back end and I also experiment with a bunch of other things like uh rust golla some assembly even objective c and while I was at the university I even invented my own programming language that nobody uses called Pomeron in Yumur style programming in your mother's tongue. Uh basically it was a Dutch translation of Pascal. Now the point is I've seen strict languages, verbose languages, functional languages, systems languages, uh scripting languages. So when I criticized Python, it's not because I don't know any of the alternatives. Now what didn't I like

about Python? The first thing that really got me with Python was that indentation defines the scope. I came from a world where basically formatting your code didn't matter because the uh syntax elements like braces defined scope. So coming from languages like C++ or Java, having to define scope by using indentation felt really fragile to me. I mean you misplace one space or you turn a tab into a space and then your logic basically changes. I was thoroughly convinced that whites space should not have semantic meaning which by the way I still kind of think but it also leads of course to an advantage which is that the code is typically shorter because you don't need to type all of these braces anymore. Now a second thing is lack of strictness in C++ or Java. Well the compiler basically stops you if you do something wrong. In Python it's possible that lots of errors only show up at runtime. Now, type annotations help, but they're optional. They're not enforced by the interpreter. You have to do work to make sure that this is taken into account. The third is that Python, in my opinion, has verbosity in really strange places. Uh, in particular, when you look at callables, if you want to specify a callable type there, the type annotations are just horrible to read. Uh, another example is generics before Python 3. 12 with the type vars. This is not great. The fourth thing that I didn't like about Python is class design where you have a dunder in it method as a constructor which always felt hacky to me. Why do I have to write self everywhere? Another thing is the whole class variables versus instance variables. This is really easy to mess up. And then in data classes you define instance variables like you would normally define class variables. It's very confusing. So compared to other languages that I use, it felt very inconsistent. Fifth thing is imports and packaging. Relative versus absolute import. When is a folder a package versus when it isn't? Why does it work in one context but not in another? It confused me for years. The sixth thing is dependency management. Virtual environments pip requirements. txt poetry pipen. Uh it's much better now fortunately with tools like uv but historically this was actually really messy. And finally, performance. It's one of the main things that I get in my comments all the time. Python is a toy language. If you want to write real code, don't use Python. Yeah, it's not fast. If you compare raw execution speed to C++, Rust, Go, it just loses. And that bothered me a bit in the beginning. Now, in the end, actually, performance is not that important for the things that I do, but still, it's in the back of your mind. So when you look at that list and actually most of those things that I didn't like about Python are still there. There's still white space that has meaning. Uh type annotations can still be verbose and a couple of other things as well. So why would I

still use Python? Now in my comment section I see all the time things like why use type hints just switch to a statically typed language? Python is a toy language. It's outdated. or from the other side, people saying that I'm some kind of Java programmer trying to turn Python into Java. Now, for the record, the last serious Java code I wrote was 25 years ago. Uh it just seems that programming languages divide people into camps. Uh I personally don't feel like I'm in a camp. I just like tools. And this is why I still use Python. It's incredibly easy to set up. It's fantastic for quick scripts and automations. It has a massive ecosystem. It integrates with basically everything, especially in AI and machine learning. Well, Python is basically the default language you should use. And as a consequence of that, AI coding agents also know Python extremely well because there's so much training data. And the big thing is that 90% of the things that I build is glue code. Connecting some APIs, getting some data from a database, orchestrating a couple of different services, integrating things like payment providers, email systems, LLM API, stuff like that. And in those cases, performance is rarely the bottleneck. The bottleneck is usually the network, the database, or an external API. So, I asked myself the question, does Python's performance matter to me? Well, most of the time it doesn't. Is it perfect? No. Is it the fastest? No. Does it solve my problems efficiently? Well, usually yes. And for me, that trade-off is just totally worth

it. The bigger lesson here, I think, is about identity. If you tie your identity to a language, then any criticism on that language is going to feel personal. Like if you say I'm a Python developer and somebody criticizes Python, that's going to feel like a personal attack. And if you do that, you stop evaluating the trade-offs objectively and you defend your camp instead of solving problems. And I think senior developers should optimize for outcomes, not ideology. There is no best language in the world. There's just the one that works best for a particular context. Even the people saying that only use Rust because it's much faster than Python. Well, Rust also has its problem, right? It's just not as easy to use as Python. So, it all depends on the context. It depends on the team, the ecosystem, the existing infrastructure that you have. If you have like a huge code base in C++, yeah, Python is great, but you're going to need to work with C++. You simply have no choice. It depends on the domain. If you need to develop a web app, you're going to use different tools. than if you need to develop some embedded software, right? And for some things like embedded systems, well, probably Python isn't the best option for that. Although I wouldn't be surprised if there is actually an ecosystem for that within Python. Uh for an AI integration startup, Python might be the best thing ever. And the nice thing about life is that you're allowed to use multiple programming languages. Yes, one project in Python, do another in Rust, do another in Typescript. Don't feel like you have to pick always the same tool. You don't have to marry your tools. And finally, don't judge other developers for their choices. You don't know what their constraints are. You don't know their team's experience level. Uh architecture, design quality matter way more than the particular language that you're using. Anyway, this video ended

up being kind of a rant. I do think it's important though to not only talk about writing code and design, but also about how we see ourselves as developers and the kind of mindset you need to develop if you want to grow into a more senior role. But I'd like to hear from you. What are your thoughts about this? Do you associate yourself with a particular language? Do you consider yourself a pythonista, a rustian, a CP? Have you ever changed your minds about a language after using it for a while? Let me know in the comments. Now, Python has its limitations, but it also has great features. If you want to learn more about those, including some features you probably don't even know exist, check out this video next. Thanks for watching and see you next