The Importance of Apprenticeship.

The "Big Four" accountancy firms have cut entry level jobs in a concerted way - KPMG cut the intake by 29% in 2024. That has gone up slightly, and they are talking about "hires which can work alongside AI" Paralegal work is another example. On LinkedIn, there are often jobs advertised where someone will help train an AI LLM to be a more convincing creative writer. 

If this is the trend across the economy, how will young people gain the necessary experience to progress in their careers? In the book 'Expert', by the surgeon Roger Kneebone, he argues that this apprentice phase is extremely important. It's an essential step on the journey to master something. He divides the phases into - apprentice : journeyman : master. 

"When you start out, your role is to learn. You are in someone else's world, learning to do things their way, and they take responsibility for what you do... You are shielded from the mistakes you will inevitably make.... You become part of a community of practice, a group of people already doing what you aspire to learn."

Large Language models are good for some uses where you are able to be approximately right. But looking at the output of something is different to having to analyse data and come to conclusions yourself. Left alone to create prompts and micro-train an LLM, how will you access the community of practice you want to be part of? If you are a writer, how will you learn dramaturgy if you are on the surface of work produced by AI. How do you know which job will still be here in fifteen year's time?

There can be productivity gains when using these tools. But the distribution of rewards is still a question nobody is sure of the answer to. At the moment, it looks like 'trickle up' economics might win out. How prolonged will the 'Engel's Pause' be, as productivity goes up but living standards stagnate, nobody can answer. 

I did a little research on resources for the Curriculum for Excellence in Scotland and was very happy to discover there were some, including this resource for teachers - trails.scot - teach-ai-literacy-handbook.

Two other things to check out if this is of interest:

Rory Stewart's series about A.I on the podcast 'The Rest is Politics' is good. The full series is for subscribers but there are 20 minute previews of the episodes available. Ep 2 is about AI and the world of work.

There is also a fascinating documentary on YouTube called 'The Thinking Game' where they tell the story of DeepMind (now Google DeepMind). There are different types of AI and this one trains itself to get better at something.

The documentary shows how they have progressed from DeepMind learning how to play PacMan, to hammering people at chess and Go, to solving protein folding (something which was in the very difficult pile). One of the founders, Demis Hassabis won the Nobel Prize for Chemistry for that. Google DeepMind also have a podcast which is good. There is no end of accessible resources.

Here is another quote from Roger Kneebone to finish.

"The opportunity to become expert should be one of our inalienable rights. It's how we fulfill our potential as human beings. Yet the slow process required is at odds with demands for instant gratification. There's a growing sense that anyone can learn to do anything - and quick. But that's not the case. Becoming an expert involves sticking at something for much longer than you normally would.

It requires...guidance and encouragement, exposures to mistakes, opportunities for spending time with your materials and the people you work with, and allowing your personality to unfold."

Let us hope the ladder isn't pulled up too quickly and that young people get their chance to do this.

END.