A simple piece of arithmetic. A child in Year 8 in 2026 will be 22 in 2036, finishing university (or not) at that point. The world they enter the workforce in is governed less by what their degree certificate says and more by what they can actually do with their judgement, their AI tools, and their relationships.

If you accept the thesis from our Khan Academy / TED-Ed piece — that the credential value of expensive degrees may erode meaningfully by 2036 — then a serious question follows: what should we, as parents, be helping a Year 8 child build now?

This article is one attempt at an answer.

The bet that's not changing

A safe-feeling bet for an independent school parent: "send my child to a good school, push for Oxbridge or Russell Group, expect the degree to do the heavy lifting on first employment." That bet has worked for several generations. It may still work in 2036 — but it's no longer the only or the best bet for most children.

The wider bet — quietly being made by the smartest independent school heads — is that the child themselves must be more impressive than the certificate. The certificate is becoming necessary but increasingly not sufficient.

That changes what we build at home. Below are eight capacities we think will matter disproportionately by 2036.

1. Judgement, not just knowledge

If knowledge is free (AI can produce it on demand), the differentiator is whether the child can evaluate it. Is this source reliable? Is this argument valid? Is this AI output correct? Should I trust this person's expertise?

How to build it now: have your Year 8 child cross-check AI outputs against primary sources. Make "check the AI's working" a normal household phrase. When the AI is wrong (which it sometimes is), don't hide that — celebrate the spotting.

2. The ability to read long, hard things slowly

The single biggest cognitive shift of the next decade is deep reading becoming rare and therefore valuable. Children whose primary intellectual diet is short-form video and AI-summarised text will be unable to sustain attention across a 400-page book or a 30-page government white paper. Those who can will be conspicuously useful.

How to build it now: protect long-form reading. Not as a chore. As a household value. Discuss what you're reading. Let your child see you struggling with a difficult book.

3. Mental arithmetic and number sense

The temptation to outsource all arithmetic to AI is real. Resist it. A child who cannot estimate, who cannot sanity-check a number, who cannot do "20% of 85" in their head, is unable to spot when the AI is wrong about a financial or business question. Number sense is judgement applied to numbers.

How to build it now: keep mental arithmetic alive in everyday conversation. Restaurant bills. Train fares. Quick estimates. The AI tutors that insist on showing working, including the mental arithmetic, are doing something durable.

4. Writing — particularly the second draft

AI can produce a competent first draft of almost anything. What AI does not currently do well is the second draft: the editing, the rewriting, the structural reshaping, the cutting of a paragraph you spent an hour on because it doesn't earn its place.

The capacity to take a draft (yours or AI's) and improve it is one of the most valuable cognitive skills you can build. It transfers to almost every professional domain.

How to build it now: have your child write something. Then have them rewrite it. Then have them rewrite it again. The struggle is the point.

5. Spoken articulation under pressure

A Year 8 child whose primary mode of communication is typing — to friends, to AI tools, to teachers via email — may arrive at age 22 unable to articulate a complex thought aloud to a senior colleague in a meeting. This is a real and rising problem.

How to build it now: make dinner-table conversation a real thing. Ask the child to defend a position. Let them be wrong. Let them rebuild. The schools that still do debate, MUN, and viva-style exams are providing a service worth paying for here.

6. Doing something physical, well

In a digitally-dominated economy, embodied skill becomes rarer and more valued. This is musical instruments at Grade 6 or above. This is competitive sport — not because they'll go pro, but because of what the years of practice did to their character. This is woodwork, electronics, hands-on engineering. This is cooking.

How to build it now: insist that something physical your child does well. It doesn't matter what. The discipline of becoming good at a physical thing is the discipline of becoming good at anything.

7. Relationships, including with adults

The economic returns to "knowing people who know things" rise sharply when knowledge itself is free. The bet your independent school is implicitly making — that the alumni network is worth something — is correct, but the network only works for the child who knows how to be in a relationship: how to ask for help without entitlement, how to give help without keeping score, how to keep up with an old friend at 30 because they kept up at 13.

How to build it now: encourage real friendships, not just classroom acquaintances. Encourage relationships with adults other than parents — coaches, tutors, family friends. Resist the pull of every child being in a screen-mediated peer-only social world.

8. Boredom tolerance

The single rarest cognitive state of 2036 will be sustained, unstimulated boredom — because every device, every app, every AI is optimising against it. The capacity to sit, undistracted, with a hard problem for forty minutes is becoming a competitive advantage of unusual size.

How to build it now: phones away during prep. Long car journeys without screens. Reading in silence. Sometimes letting the child be bored, with nothing to do. It is uncomfortable. It is also a gift.

The role of AI tutoring, honestly stated

Where does an AI tutor fit into this list? Most directly on items 1 (judgement — checking AI outputs), 3 (mental arithmetic, if the tutor refuses to do it), 4 (writing — though we don't do English yet), and 8 (boredom tolerance — only if used in disciplined short sessions, not as endless entertainment).

AI tutors badly used corrode several of these. AI tutors well used reinforce them. The difference is whether the AI insists on the child doing the cognitive work, or quietly does the work for them.

This is the central design choice at aitutors.me. Professor Pi refuses to give the answer. The Socratic ladder forces the child to think first, ask second. The energy gating insists that exhaustion is real and tutoring through it is counter-productive.

It is not a perfect design. But it is a design oriented at the eight capacities above, rather than against them.

What about university itself?

A direct question many independent school parents are quietly asking: should my child still go to university?

Our view: yes, in 2036, for most academically able children, still yes — but with eyes open. The thing university increasingly provides is the community (friends for life), the networks (the people you'll work with at 30), the transition (the supervised independence of 18–22), and the brand (which still matters at first hiring). What it provides less of, by 2036, is content. The content is free.

Choose the university for those four things, not for the content. And help your child arrive at 18 with the eight capacities above already partly built. The combination produces a 22-year-old in 2036 who is meaningfully more able than the typical graduate of 2026.

A closing line

The job of education has not changed: it is to make a child into an adult who can act in the world with intelligence and judgement. The tools have changed. The job has not.

The children in Year 8 in 2026 are not at a disadvantage. They have something previous generations did not: an AI capable of running a one-to-one Socratic conversation with them every evening for the price of a takeaway. Used badly, that destroys the eight capacities above. Used well, it accelerates all of them.

The choice is ours.

Jason runs aitutors.me, has a daughter in Year 8 at a UK independent day school, and writes the household policy on screen time from rough conviction rather than research. Updated 21 May 2026.