What ChatGPT and My 2-Year-Old Niece Have in Common (Spoiler: A Lot)

We’ve been experimenting with an AI-generated ‘digital presenter’ to dive into how to think about AI with you. Take a look and see what this AI-powered approach can do.

👇Watch: Check out how our AI presenter can bring you a shortcut to our blog post. Instead of reading, you can watch our blog presented to you in a short, 3-minute video.

In today’s email:

• What ChatGPT and My 2-Year-Old Niece Have in Common

• Bits and Bots: Introducing our AI-Generated Presenter

• What I’m Consuming: Interview with Sal Khan (Revolutionising Education with AI)

What ChatGPT and My 2-Year-Old Niece Have in Common (Spoiler: A Lot)

I was hanging out with my two-year-old niece recently and realised she’s like a tiny sponge, absorbing everything around her—words, gestures, and the occasional bad habit. This got me thinking: Large Language Models (LLMs) are a lot like two-year-olds on a data binge.

A Large Language Model (LLM) is a type of artificial intelligence trained on vast amounts of text to generate human-like responses, much like a child mimics what they see and hear.

Observing My Niece’s Learning Journey

The World Through a Two-Year-Old’s Eyes

My niece learns from her parents, kindergarten, and pretty much anyone she encounters. She mimics what she observes. If her parents always say “please” and “thank you,” she picks up on being polite. But if she hears others using inappropriate language, she’ll copy that too.

How Children Learn Language and Behaviour

Children naturally imitate what they hear and see. At kindergarten, my niece learns new words and catchy songs. She even invents her own words—like calling broccoli “tree snacks.”

For example:

• My niece calls every dog a "woofie," just like her parents do.

• Living in Amsterdam, she mixes Dutch into her Afrikaans, thanks to the Dutch spoken at her daycare—even picking up a little Dutch accent.

• Whenever she does something funny, she’ll exclaim "silly billy" or "grapjas," mixing her languages effortlessly.

• During nappy changes, she cheerfully says "bingo, it’s a mingo!"—a line from her dad, inspired by the flamingo on her nappy.

• She turns off lights and shuts doors, imitating her dad’s neat habits.

• She adorably calls smoothies "foodies" and cottage cheese "cockachee."

An Epiphany: Connecting to Large Language Models

Drawing Parallels Between Children and AI

Watching my niece learn reminded me of how LLMs absorb information in a similar way. Just as my niece learns from everything around her, LLMs process vast amounts of text data to “learn” how to generate language.

Continue to read the full blog post

Bits and Bots (and some fun)

What I’m Consuming:

Interview with Sal Khan: Revolutionising Education with AI

Sal Khan, founder of Khan Academy, discusses how AI is transforming education, particularly for on-the-job training.

Key points include:

• Khan Academy's Mission: A non-profit providing free, world-class education through personalised exercises, teacher tools, and videos.

• AI’s Role in Learning: AI makes education more accessible, responsive, and personalised. Generative AI allows learners to ask follow-up questions and explore specific scenarios after watching training videos.

• Upskilling & Lifelong Learning: AI tools like Khanmigo help people deepen their understanding of various fields and emphasise the importance of continuous learning.

• Soft Skills Matter: Beyond traditional skills, AI can help develop communication, collaboration, and empathy—essential skills in the modern workplace.

Here’s a key takeaway:

If you forget everything else, remember this…