
Advice for Junior Developers

AI is Smart Until It's Dumb: Why LLM Will Fail When You Least Expect It - Emmanuel Maggiori
Topic 15 of 19
Published Nov 10, 2025 · Clip 56:49–59:35 (2:46)
For juniors, Emmanuel advises working on complicated, tightly specified, custom work that AI can't replace because you can't even describe it in a prompt—like his project modeling heat propagation in buildings with physical models and API-connected devices. Conversely, MVP/prototype/no-code-style work is most at risk. He also recommends mission-critical software—where mistakes lose money or harm people—since it demands manual validation AI won't replace soon.