Sitting together with many parents and kids
We just sat together and talked … about AI, what else? Everyone in the group is using it and everyone somehow copes with the crazy flood of new foundation models, tools, approaches, or just news, or own experiences in this field.
It seemed that everyone in the group has a more or less well working mental model around LLMs and everyone is more or less able to use things like ChatGPT well enough and fearless enough to get to good enough results through experimentation and experience build-up. On the side, the tools and models are getting better anyway at “guessing” what we want them to do and LLM companies also invest in solving known issues better with each new generation of their products.
One set of questions popped up that somehow challenged our approach so far: Are we actually able to explain to our kids how AI works? How well do our mental models about AI and LLMs really work? Will we be able to keep up with the rapid developments in the field? How often will we invest time to wade through explainer videos on youtube to understand new developments?
At the same time: Will this be really necessary? There have been other - sometimes disruptive - everyday systems which are not meant to be understood by non-expert users: computers, car engines, financial systems - including money, our stereo system or TV. Kids proved to be “native” adopters of all of these things. There was never a need to eplain the 5 year old how the iPad works. That being said, I doubt that kids will really come asking for explanations how AI actually works. I think it’s rather our responsibilty as parents to maintain at least a basic level of AI literacy and empower our kids to responsibly use AI to their advantage.
We ended up with the idea that I would at least collect a list of not too hard to digest explainer videos and go-to places to reduce the suffering in case we wanted to look inside the enclosure.
So without further ado, here’s our curated list…
3Blue1Brown
Excels at visualization: Great diagrams and animated graphics to explain neural networks, attention mechanism and the like. I recommend watching the playlist start to end and skip videos that are not interesting enough. I think the content becomes harder to understand towards the end of the list.
The Art of the Problem
Starts from history of Deep Learning and Machine Learning and provides some deep dives into recent developments. I would start with the video on “What is Machine Learning?”
Deeplearning.AI
Company founded by Andrew Ng. Lots of free courses. Provides opportunities to dive deeper and implement something on your own.
Andrej Karpathy
It is really pleasant and entertaining to follow his deep dives and explanations. Andrej is a OpenAI employee. In his videos he talks about ChatGPT inner workings and he builds software on the go, and the viewer can play along.
I would start with this video: Deep Dive into LLMs like ChatGPT
Checker Welt (german)
If you and/or your kids understand german, this is the right place for kids to learn something about AI themselves. It doesn’t go to deep into the current LLM/Agent hype and seems almost timeless. Unfortunately, the content of the english videos I came across delivered rather shallow content.