`# 👋 Hi there!

I'm Bojan (🔊), but if that’s hard to pronounce, you can call me Ben — it’s close enough. 😉

• Location: Belgrade, Serbia, South-Eastern Europe.
• Background: IT & FOSS enthusiast, self-taught programmer, and long-time Linux user

• C • C++ • POSIX shell • Python • Perl • Pascal • BASIC • Ruby • PHP • SQL • a bit of Assembly (x86, 6502, 8080)

• Rust, Go, Nim, Zig, Lua, Scheme, Lisp, assembly (ARM7, ARM64).

• Main OS: Slackware Linux (since 1999), Current.
• Other OSes: macOS Tahoe (26.1), Windows 11.
• Retro/Historical: DOS variants, UNIX and adjacent systems, old Linux and ELKS, CP/M, OS/2, Windows 1.0-9x/NT, AmigaOS, Atari TOS, classic MacOS, Mac OS X.
• Alternative: HaikuOS, AROS, ReactOS.

• Retro computers: Commodore 64, Commodore 128, Atari ST, Amiga 500, ZX Spectrum 48k.
• Embedded/SBCs: Planning work with: RPi/Arduino (or alternatives) for IOT/gardening/permaculture projects. Interested in building a custom distro with Buildroot/Yocto.
• Other interests: embedded Linux, system engineering, system administration, and hardware and software preservation

• Music: guitar & bass (blues, soul, rock, funk, early R’n’B)
• Drawing: mostly cartoons and caricatures
• Writing: drafting a fantasy web novel
• Nature: relaxing near water, walking in nature
• People: spending time with friends and family
• Pets: currently don't have one, but thinking of getting a cat.

• English – Near-native
• Italian – Read comfortably, speak at a basic level
• French – Can read well, speak a little

• Chinese • Japanese • Spanish • Portuguese • Greek • Polish • Russian

• Moderator & contributor at MyCity.rs forums
• Admin at Slackware Serbia community
• Occasional contributor to SlackBuilds.org
• Occasional small contributions to various FOSS projects (nothing major so far)

• CompTIA A+
• CompTIA Network+
• CompTIA Linux+
• LPIC-1

• Worked with a wide range of operating systems, including multiple Linux and BSD variants
• Packaged software for Ubuntu, Fedora, Slackware, and AppImage
• Created minimal chroot-based Linux environments for experimentation
• Used various virtualization solutions on Linux (QEMU/KVM, VirtualBox, etc.)
• Familiar with multiple Linux/BSD init systems and package management tools
• Experience with chroot and LXC containers
• Some hands-on experience with CI/CD pipelines
• Performed LAN setup, maintenance, and troubleshooting
• Provided IT equipment procurement and consulting
• Diagnosed and resolved hardware and software issues in a small company
• Maintained and administered a community website on shared hosting
• Served as a forum moderator and mentor, helping users with technical and community issues
• Wrote and translated Linux-related tutorial articles
• Some public speaking experience, though still not confident in that role

• Over the past year I’ve been experimenting with AI-driven development. It’s a long topic and I have plenty of thoughts about it, but here’s the short version: AI isn’t replacing developers anytime soon, but it can be genuinely helpful, especially on smaller, well-defined tasks.

• To test this properly, I tried building a bunch of different things with AI support:

  • text editors (console and GUI),
  • file managers,
  • a man-page viewer,
  • a small Markdown authoring tool,
  • simple lexers, parsers, interpreters,
  • a few other utilities.

• I also experimented with using AI to:

  • debug code,
  • add new features,
  • adapt existing projects,
  • and even generate shell scripts (which works, but you really have to double-check everything).

• Some of these projects might get released at some point once I clean them up. The basic code is there; it just needs some shaping before it’s ready for the world.

• One thing I noticed is that AI’s usefulness drops as a project grows in size or complexity. At that point you basically switch back to human-led development and use the AI as a helper, someone to ask for a quick review, bounce ideas off, or help track down an odd bug. AI is also useful when exploring feature ideas, but only if you already know what you want.

• So here’s my takeaway:

  • AI as a fully automatic programmer: not reliable.
  • AI as a helper for small tasks: very good.
  • AI as a learning tool: surprisingly effective, especially if you already know the basics and ask the right questions.

• 2025.11.08. - I actually broke GitHub stars by having more than 5k stars. I had to remove some to access the rest. For the more important ones, I ended up either saving the content locally on the disk or documenting them via Zread.ai. 5k is too much to manage manually, so I don't know what to do except to ocassionaly remove some in passing.
• 2025.12.02. - I cleaned up a lot of starred repoes. I have around 9.980 left (still a lot). It's likely that number was at least 500-1000 starred repoes higher before. I also deleted all groups. Now I can access "the stars" much more easier and rarely see the unicorn again.

⚙️ Always curious about how things work, how to break them and then fix them again.