I originally posted this on Mastodon, but I thought I'd add it here too: “What I mean is that if you really want to understand something, the best way is to try and explain it to someone else. That forces you to sort it out in your own mind. And the more slow and dim-witted…

NeilMadden's blog is a resource for developers and software engineers seeking to enhance the security and efficiency of their programs. Covering topics such as secure coding practices, encryption techniques, and vulnerability management strategies, it offers practical insights and actionable advice to mitigate security risks effectively. Through detailed tutorials, code examples, and expert commentary, the blog equips its audience with the knowledge and tools needed to build robust and resilient software solutions in today's digital landscape.

Neil Madden

A brief personal stance on why LLMs shouldn't replace the act of programming, drawing on quotes from Douglas Adams, Carl Friedrich Gauss, and Alan Perlis. The core argument is that programming is fundamentally a learning and thinking process — the act of breaking down problems for a machine teaches the programmer. Outsourcing that to an LLM bypasses the cognitive work that generates real understanding and satisfaction.

Why I don’t use LLMs for programming

These posts all miss the fundamental point. AI has essentially forced all developers to become “managers” (but of AI, instead of junior devs or of themselves/their own code) - whether they like it or not - and some aren’t ready or equipped for that change.
The delegation process is itself a skill - just a different one - which, judging by the numbers of these posts, is something devs have yet to adapt to.

I am also sure that developers will become worse at their job when they stop practicing. In other words, when they stop writing code. Reading code is not the same as writing and thinking. There is scientific proof about how writing enables learning. Programming is like using a language. When you stop practicing, you’ll first start to struggle and eventually keep only small crumbs that are not enough to speak and understand properly.
AI coding is based on an ambiguous language for a non-deterministic probability machine. It will always make logic errors, because it doesn’t understand what it does. As a codebase grows these issues will accumulate and result in a no longer maintainable mess. The problem becomes bigger the more you entirely rely on AI and give up your coding abilities and thinking.
I don’t care if that’s the case for a non-critical short lived application like a website. But I care for mission critical software that should support a company for at least 10 years.
In my opinion the best approach is to split your work in multiple small jobs and hand over the boring stuff to the AI, while you also write code and think about what you’re doing, while you’re refactoring and improving the initial approach. This gives you the best out of two worlds. Let the AI only write code that you can understand and that’s small enough so you can review it later with the help of another AI. But never stop writing code yourself! Don’t give up the beautiful craftsmanship you have worked on for many years.
This is also the reason why I write all my comments and the majority of my articles without the AI. I want to practice and relate to what I say and write.

Yeah, please enlighten us how you’re using candles in your house also and only eat what you grow in your garden…
With the help of AI I do in one day what will take me 2 weeks and it is my duty to my employer to deliver ASAP and in good quality (which I certainly do since I review everything).
If you choose to NOT use AI for coding, simply to retain your coding abilities, for the sake of “fun” then I get it, but having a job demands us to use the best available tools to complete the tasks at hand ASAP or else they will quickly find someone else who will.

Skill issue.
Using AI ≠ Incompetence