June 15, 2026
Render me shocked
Show HN: I wrote a C++ ray tracer from scratch without AI
Coder says he built stunning image software solo, and the internet is obsessed with the 'no AI' flex
TLDR: A developer spent years building a realistic image-making program from scratch, and the internet immediately fixated on the bold “without AI” claim. The comments turned that phrase into the real event: part joke, part status symbol, and a surprisingly heartfelt celebration of old-school coding.
A developer dropped a passion project called Luz, a home-built image renderer that simulates light, shadows, reflections, fog, and camera effects to create realistic pictures. On paper, it’s a giant technical flex: years of work, no outside code libraries, works on Mac, Windows, and Linux, plus a long menu of features for artists and tinkerers. But let’s be honest: the community instantly locked onto the spicy part of the title — “without AI.” That phrase basically stole the show.
One commenter joked that “without AI” is the new version of bragging that something was “written in Rust,” turning the post into a mini roast of modern tech marketing. Another laughed, “Now this is how you catch attention,” which is pretty much the mood of the whole thread: half admiration, half wink at the headline game. The subtext was loud: in 2026, saying you made something without AI is its own badge of honor — and maybe its own meme.
Still, there was real love underneath the jokes. People shared that building a ray tracer was a beloved school project and a rite of passage for graphics nerds, with one commenter saying they did a simpler version back in 2016 and loved it. Others jumped in with helpful recommendations like The Ray Tracer Challenge for anyone tempted to try it. So yes, the comments came for the anti-AI drama, but they stayed for the nostalgia, respect, and “wow, this person really built the whole thing by hand” energy.
Key Points
- •Luz is a C++20 path tracer built from scratch with no third-party dependencies.
- •The renderer supports Monte Carlo path tracing, global illumination, multithreaded CPU rendering, adaptive sampling, denoising, atmospheric scattering, and BVH acceleration.
- •It includes custom .luz scene files, a Blender-to-Luz conversion workflow, support for multiple geometry and material types, and BMP/TIFF output.
- •The project provides Makefile and CMake build paths, a test suite, CLI-configurable render settings, and deterministic benchmarking tools.
- •Luz supports macOS, Linux, and Windows, and its release builds use aggressive optimizations such as -O3, -march=native, and LTO by default, with documented options to disable them.