$ cat /var/log/blog.log

About a year ago I switched from arch linux to NixOS as a daily driver on my laptop... I fell in love with the idea of NixOS: as input you (basically) define your whole system configuration in a special config file syntax (Nix) and as output it gives you a reproducable build of your whole operating system and keeps track of changes so you can always revert back to an older generation if necessary. While this is nice on paper, it didn't end up working out for me.. After using NixOS for almost 1 year I decided to switch back to Arch linux on my lapop. Here's why: ## Nix break all the Time.... Ironically, I only broke Arch once in five years, whereas NixOS often breaks even before updating. Something usually requires a change in my Nix configuration before the `nixos-rebuild` command finally succeeds. It's good that this check exists, but the constant cycle of `rebuild → fix → rebuild → fix → rebuild` becomes quite annoying after a while especially when the required changes are very small... When the new configuration finally builds, more ofen than not some component randomly stops working after reboot. If it isn't broken copy paste between electron apps it is audio or bluetooth that randomly stopped working. I'm left alone to fix it with 0 clue as to why it broke. ## Huge update sizes NixOS handles dependencies differently compared to Arch Linux, which leads to much larger update sizes. Instead of replacing packages in place, NixOS installs new versions alongside the old ones, keeping multiple system generations.... For example, updating `glibc` in Arch Linux simply replaces the previous version. In NixOS, every package depending on glibc gets rebuilt or redownloaded, which greatly increases disk usage. While this provides rollback safety, it also means the system can quickly grow in size (unless you run **nix-collect-garbage** periodically). ``` $ ls /nix/store | grep glibc-locales 0j3cpwmwab04jmvyzcqh538jzrnm60hn-glibc-locales-2.40-36.drv 0k4azawc10ygacxyqjz36cqghxsicqpr-glibc-locales-2.40-66.drv 2i4wm7zsm2nkn7sn2aj7ng4x1k3p71gz-glibc-locales-2.39-52.drv 2sk3c2n6p3aavhd7659ll2w63mv0i3aw-glibc-locales-2.40-66.drv 33alnvvkrp1s13vg5frn1zca34xbz424-glibc-locales-2.39-52.drv 51ssq420b3wqvcxmgmcw00hk8zkh6gvw-glibc-locales-2.39-52.drv ... ``` If I take this, Arch Linux updates are easier to manage. Old files are removed automatically, and shared libraries don’t trigger redundant downloads. If you care about minimizing disk usage or have limited internet/network, NixOS can be a nightmare. ## Compilation takes forever (and binary caches are unreliable) NixOS builds packages in isolated environments, which often means compiling everything from source even for minor updates. While binary caches like [Cachix](https://www.cachix.org/) (a hosted Nix binary cache service) can theoretically speed up installations, they often miss packages due to system differences, forcing unnecessary compilation. For example, if a dependency isn’t cached exactly as required, NixOS will rebuild it locally even for common packages. This can take hours, especially on slower hardware. On my machine, regular maintenance updates without proper caching easily take 4–5+ hours.... In contrast, Arch Linux simply downloads prebuilt binaries via pacman or an AUR helper, making updates take just a few minutes. If you dislike waiting for your system to rebuild constantly, NixOS can be frustrating. And if you actually enjoy recompiling everything from scratch, why not use [Gentoo Linux](https://www.gentoo.org/) instead? ## Conclusion I wish NixOS a bright future ahead to address all of its shortcomings. But until then it's no longer my favorite distro and I wouldn't recommend daily driving it as your desktop linux distro unless you have a very specific use case.... :::note I do not hate NixOS, but it’s not really built for me. I prefer simplicity and speed over complex builds. :::

Artificial intelligence is reshaping industries at a pace rarely seen in modern technology. As organizations accelerate their adoption of AI systems, concerns about job displacement have grown louder. However, the underlying reality is more complex: AI is not replacing humans as much as it is redefining how human work is structured and performed. In this context, AI refers to technologies such as large language models, machine learning systems, and generative tools that automate or augment tasks traditionally performed by people. These systems are designed to optimize workflows, not replicate the full scope of human intelligence. ## Tools as Multipliers of Human Capability Throughout history, technological tools have functioned less as replacements and more as multipliers of human effort. The introduction of industrial machinery did not eliminate skilled labor; it increased its efficiency and scale. AI follows the same pattern. AI systems are particularly effective at handling repetitive tasks, processing large volumes of data, and generating outputs at speed. Yet they lack qualities that remain distinctly human, including contextual judgment, emotional intelligence, ethical reasoning, and creative intent. ## AI as Infrastructure, Not Authority Across multiple sectors, AI is increasingly positioned as an assistive layer rather than a decision-maker. - **Creative industries** use AI to generate drafts and concepts, while humans provide narrative direction, cultural understanding, and originality. - **Healthcare systems** deploy AI for diagnostics and pattern recognition, but clinicians retain responsibility for interpretation and patient care. - **Software development** integrates AI for code generation and debugging, while engineers continue to design architectures and solve complex problems. In each case, AI functions as infrastructure that supports human decision-making rather than replacing it. ## Adapting to an AI-Driven Economy The growing integration of AI into professional environments is shifting the skills that matter most. Three trends are emerging: 1. **Continuous learning** is becoming essential as tools evolve rapidly. 2. **Human-centered skills**, such as critical thinking and creativity, are increasing in value. 3. **Ethical oversight** remains necessary to ensure AI systems are deployed responsibly. Resistance to AI adoption is unlikely to prevent change. Instead, it may widen the gap between those who leverage new tools and those who do not. ## Conclusion AI is not a force that eliminates human relevance; it is a system that amplifies human capability. The future of work is less about competition between humans and machines and more about collaboration between them. As AI becomes embedded in digital infrastructure, the most significant advantage will belong to individuals and organizations that understand how to integrate these tools effectively while preserving human judgment at the center of decision-making.