Three sentences — a style name, a push-the-limits intensity cue and two technical constraints — are enough because neobrutalism is a visual language the model already has abundant training data for.
View original source ↗Prompt
Make a neobrutalist webpage, make it extremely creative, as far as possible, push the limits. Add smooth scroll animations, add fancy colors and tailwind css styles. Make it responsive. title of the page is dorksense
Reusable method
Key decisions
Check whether information hierarchy, module order, and CTA clarity all hold in the final result.
Adapt it to your subject
Extract “target result → the product name, audience, page modules, and primary button → constraints” as a reusable template for similar work.
Where it breaks
Run single-variable comparisons around module order, value proposition, and CTA placement. Record the model, cost, and failures; turn it into a Skill only after it repeats reliably.
Recommended models
Creator
Dorksense focuses on production-ready AI interfaces and is active on X. “One-Line Neobrutalist Webpage” is the current representative case, built with Gemini 3 Pro with its stability retest open to votes.
Keep exploring
A web case by @chetaslua created with Gemini 3 Pro, including the public result, full prompt, and original source.
Prompt
🚨 Gemini 3.0 pro - ecpt checkpoint Holy shit , our prompt and imagination is the only limit check the website Prompt : Write code for a mysterious website about simulation theory. Make it feel like reality is rend
A three-line tech-stack list doubles as the implementation plan — camera feed, Mediapipe landmarks, Three.js rendering — with a clear data flow between each step instead of a description of the visual effect.
Prompt
Use the computer's front-facing camera Use `mediapipe` lib to capture facial landmarks Use `threejs` to apply a LASER EYE effect to the face captured by the camera based on the real-time 3D landmar
Style information lives entirely in an uploaded reference image; the text only names the task and the output format, testing cross-modal image-to-code ability rather than prompt-writing skill.
Prompt
I have provided an image. Code a beautiful voxel art scene inspired by this image. Write threejs code as a single-page
