📣 Learn platform feedback — tell us what's confusing, broken, or brilliant

**[gpu.rocks/learn](https://gpu.rocks/learn)** is an interactive GPGPU course — 4 tracks, 21 modules and 107 hands-on tasks that run on your actual GPU, right in the browser.

This issue is the front door for feedback of any kind. Comment below with whatever you've got:

- **Stuck or confused?** Tell us the task (paste its link, e.g. [Grayscale, the GPU way](https://gpu.rocks/learn/data-in-data-out-42b68d01/3)) and what you expected a hint to say.
- **Something broken?** Include the task link, what happened, your browser, OS and GPU, and the mode you ran (Auto / CPU / GPU).
- **Benchmark surprises?** GPU vs CPU results vary a lot by hardware — we'd love real-world numbers either way. Some tasks are deliberately faster on the CPU, and knowing where the crossover sits on your machine is genuinely useful.
- **Ideas?** Missing topics, a track you wish existed, editor ergonomics — all fair game.

> Your progress and code never leave your browser (localStorage only), so paste your kernel code into a comment if it helps us reproduce something.

Clear reproducible bugs may get promoted into their own issues so they're easier to track. Thanks for helping make GPGPU easier to learn 🪼

Start at gpu.rocks/learn and choose a specific task or track concern to reproduce in the browser. The issue names no repository files or tests; a useful outcome is a focused report with the task link, expected and actual behavior, browser, OS, GPU, run mode, and any relevant kernel code.

Written by the indexing model from the issue text.