Use AI for speed. Keep your own judgment sharp.
AI can draft code quickly, but you still need the instincts to debug, question, simplify, and own the final result.
How do you stay sharp while coding with AI assistants?
To stay sharp while coding with AI, keep doing small implementation reps without outsourcing every step. EraCode helps by giving you stack-aware challenges and feedback that exercise reasoning, not just prompt writing.
AI changes where weakness hides
Before AI, a gap in understanding usually slowed you down immediately. Now the code may appear before the understanding does.
That makes practice more important, not less. You need enough fluency to inspect generated work, adapt it to your codebase, and catch the parts that look plausible but are wrong.
Faye’s synthesis of Anthropic’s workplace research highlights a sharp version of the problem: supervision and debugging judgment are exactly the skills that can erode when assistants remove the reps that used to build them—a tension summarized in Agentic Coding is a Trap.
Practice the parts AI does not own
EraCode challenges ask you to reason through code, tests, tradeoffs, and feedback. The reps are intentionally small so you can keep the habit while still using modern tools at work.
The point is not nostalgia for manual typing. It is maintaining the judgment that makes AI useful instead of dangerous.
Good to know
EraCode is complementary to AI coding assistants. It is not trying to be the assistant you use to write production code.
When a challenge is timed, we use a server-anchored timer and combine your AI score with how long you took—across coding, terminal, and multi-part submissions.