To learn algorithms, first draw the structure, simulate the process, and explain the invariant; only then turn the idea into code.

Auto-scroll in an AI chat is not just scrollTop. It needs intent signals, a follow/free state machine, handling for programmatic scroll events, virtualization, and mobile edge cases.

Based on hands-on use of Copilot, Cursor, and Claude Code, this post separates where AI assistants help most, such as search, boilerplate, and error diagnosis, from where they still struggle, such as architecture and edge cases.

After completing BlueDot Impact AI Safety Fundamentals and receiving the certificate, I have a clearer view of agent safety evals, tool-use safety, and reliable AI infrastructure.