How fork() duplicates a process without copying its memory

I made a visual explainer on how copy-on-write works in Linux.

When a 10 GB process calls fork(), Linux does not immediately copy 10 GB of memory. It duplicates the page tables, points both processes at the same physical pages, marks them read-only, and waits for the first write.

The video also covers things like page faults,exec(), Redis snapshots, Android Zygote, lazy zero pages, memory overcommit, COW storms, and some CVEs, etc