Thinking, empathy, and human value in the AI era

Executive overview

AI can now solve creative problems, write fluently, and outperform most humans on analytical tasks. The one thing it cannot replace is a human who genuinely wants to create value for others.

The skill that matters most going forward is not creativity or technical ability — it is thoughtfulness: the capacity to reason independently, simulate the world from another's perspective, and care about the outcome for other people.

The person who thrives in the AI era is one who thinks for themselves, builds trust, and is motivated by making others better off.

Why writing and language still matter

• AI is a large language model — its power is the L. If humans stop writing, they stop thinking.
• Using AI to do school writing is like driving a car for exercise: you get none of the benefit.
• Reading, writing, logic, and communication are the foundations of independent reasoning.
• The ability to think through unfamiliar situations is more valuable than any memorised technique.

What AI cannot replicate

• AI can now solve International Math Olympiad problems — creativity is no longer a human moat.
• The only remaining human edge: we care whether other humans survive and thrive.
• The only sustainable reason people will want you on their team is that you authentically create value for them.
• If you cannot simulate the world through another person's eyes, you cannot identify real problems worth solving.

How to interview for real intelligence

• Ask questions until body language shows the candidate has never seen the problem before.
• Give new hints they would not have encountered in school.
• Observe how quickly they synthesise unfamiliar ideas into a working solution.
• That synthesis speed is the signal — not prior knowledge or polished answers.

The danger of outsourcing thought

• Reliance on AI for reasoning produces adults who accept whatever they are given.
• Attention spans are already eroding from device use; AI accelerates this.
• A population that cannot reason is easy to deceive — including by selective but technically true stories.
• Every information source has an agenda; you can only detect it if you can think independently.

How to use AI well

• Use it to build understanding inside your own brain, not to produce output on your behalf.
• Cross-check AI across multiple tools and sources; each has embedded bias.
• Consume deliberately biased sources from opposite ends to see where they diverge — that gap reveals the real dispute.
• AI will sound so complete on contested topics that you may not notice what is missing.

The Live program: scaling critical thinking

• High schoolers aged 15–18 coach middle schoolers aged 10–13 in live, interactive math sessions.
• Professional actors and comedians train the high schoolers in charisma, communication, and presence.
• High schoolers gain the non-technical skills — confidence, leadership, persuasion — that will actually limit their careers.
• Pairs of coaches teach together, building a trusted network of kind, clever people who know each other.
• The goal is a resilient community capable of responding to civilisation-scale challenges.

Social entrepreneurship that sustains itself

• A good social enterprise must make a net profit on every person it helps — not just do good on donations.
• Start by identifying what the unsolved problem actually costs the world, then find who is already paying that cost.
• The Live model: paying customers fund enrichment classes; profits subsidise students who cannot pay.
• It took eight years to arrive at the actor-plus-math-coach model; most ideas were discarded along the way.

The idea generation method

• Generate with excitement, then immediately try to destroy the idea with full effort.
• 99% of new ideas are fundamentally flawed; hunting for the flaw is the work.
• Keep asking "is there another way?" as a default habit, not only when stuck.
• The 1% that survives rigorous destruction is worth pushing hard.