In software, we talk constantly about learning — new languages, new frameworks, new tools. We talk much less about the two practices that actually keep experienced engineers sharp: unlearning what has become wrong, and relearning what we only half understood the first time.
Unlearning
Unlearning starts with an uncomfortable admission: some of what you know is incomplete, outdated or simply false. The practice is to recognise the bias in your existing knowledge, remove what no longer holds, and start again with a clean slate.
The signals are recognisable. Your methods feel outdated. You struggle to keep up with changes in the industry. Feedback keeps revealing gaps you didn't expect. Or — the most reliable tell — you notice yourself resisting new approaches rather than evaluating them.
None of this is easy. Letting go of long-held beliefs produces genuine cognitive dissonance, and the period between dropping the old understanding and building the new one is uncertain and uncomfortable. That discomfort is not a sign you are doing it wrong; it is what unlearning feels like.
Relearning
Relearning is different: it is revisiting a concept after gaining experience. You understood it once — or thought you did — and now practice has exposed how superficial that understanding was. Each revisit deepens it.
The signals here: you can explain a concept but not fully apply it. New contexts shed light on gaps. A new technology forces you back to fundamentals you skimmed. The cycle is simple — gain an initial understanding, apply it, then return to the concept with your experience in hand.
The main obstacle to relearning is the belief that you already know. It takes some humility to reopen a topic you consider finished, and some patience to accept that understanding deepens slowly. Finding the time is a challenge too — revisiting old ground never feels as urgent as learning something new.
The mindset underneath
Both practices rest on the same foundation: treating your current knowledge as a working draft rather than a finished document. Regularly audit what you know for obsolescence. Stay curious. Treat failure as information. Apply new knowledge in real work rather than letting it stay theoretical.
Skills age fast in this industry. The engineers who stay relevant are rarely the ones who learned the most at some point — they are the ones still willing to unlearn and relearn today.