The tooling landscape settled: a few categories genuinely accelerate developers; the rest is churn. Here is the useful map for students.
The categories that matter
- In-editor autocomplete (GitHub Copilot): free for students via the GitHub Student Pack — completes boilerplate, tests, repetitive patterns. Biggest single productivity add
- Chat assistants (Claude, ChatGPT): explain concepts, review code, debug errors — the personal-tutor use case (see how to use AI without wrecking your skills)
- AI-first editors (Cursor, Windsurf): chat-with-your-codebase, multi-file edits. Popular in startups; free tiers exist. Learn plain VS Code first — same editor core
- UI generators (v0, etc.): prompt → React component. Fine for scaffolding; you still need to understand and modify the output
What interviews still test (unchanged)
Whiteboard/shared-editor coding with AI OFF, explaining YOUR project's decisions, debugging unfamiliar code, fundamentals. AI moved the bar: typing speed matters less, understanding and verification matter more — you must be the person who catches the AI's wrong answer.
Student action list
- Claim the GitHub Student Pack (free Copilot + credits)
- Use AI as tutor/reviewer daily; code solo on the judge weekly
- Mention AI-augmented workflow in interviews with a concrete example — it's a plus when paired with proof of solo skill