Ask someone under thirty to memorise ten phone numbers and they may look at you as though you proposed sending a fax. We did not become less intelligent; the environment stopped rewarding that particular skill.

AI will produce similar shifts at larger scale. Drafting routine emails, creating first-pass presentations, summarising reports, translating, coding common functions—skills that once signalled competence may become cheap.

The obvious gain is productivity. The less obvious question is what happens to the learning ladder. Senior judgement is often built on years of doing junior work. If AI removes the junior tasks, organisations must find new ways to create the underlying mental models.

Consider writing. Struggling to organise a paragraph is not wasted time; it can be the process through which thinking becomes clearer. If the paragraph arrives instantly, the output may improve before the writer does.

This does not mean preserving obsolete work for nostalgia. Nobody should calculate payroll manually to build character. The task is to identify foundational skills that remain necessary for supervision, verification and original thinking.

Education may need to shift from producing answers to defending them. Students can use AI, but must explain assumptions, critique outputs, compare alternatives and demonstrate independent understanding.

Professionals can protect skill by occasionally working 'AI-off' in areas they want to retain. A pilot practises manual procedures despite automation. Knowledge workers may need equivalent drills.

Technologies always make some skills unnecessary. The important question is whether we intentionally choose what to let go—or discover too late that we outsourced something we still needed.