Memorisation Still Matters in the Age of ChatGPT

Information is instantly accessible and with the coming of ChatGPT, now everyting is instantly understandable. I can type "Explain quantum physics to me like I'm a five year old" and I get back a cohesive answer. Is it correct? I don't know - who cares!

Since we have instant access to information we can instantly understand, can't we remove the need for memorising? We never liked it in school anyway. What was the point of learning multiplication tables or essays about historical events except to regurgitate it in an exam? These poisonous memories of memorizing unimportant information surely proves we don't need to memorise anything!

Let's look how the brain works:

Working memory is the stuff that we're consciously thinking about right now. Our working memory receives information from the environment and interacts with our long-term memory. When information isn't available in our long-term memory, or we can't 'reach' it, we need to find an answer, so we might look it up online. This is an example of an "External Resource".

In essence our working memory pulls information from the environment and long-term memory. If something isn't available in long-term memory, we need to find it somewhere else.

For example, if I'm in a quiz and get asked the question "What year did world war two end?", I can either access my long-term memory to pull that information, or I can sneak my phone out under the desk to look it up.

Advantages of Long-term memory over outsourcing.

Working memory always attempts to access long-term memory first. If it can't find anything (ie. we're not sure about the answer or how to solve the problem) then we need to get the information elsewhere. So we go externally.

Having that knowledge stored and available in our long-term memory for instant access acts as a multiplier on our ability to think and solve problems fast. If I have the information in my long-term memory (ie. I've learned it previously), then I can answer the question or problem immediately.

Moreover, knowledge is not stored in a vacuum, but in a web. Each piece of knowledge is connected to another piece of knowledge. So the more I know, the faster I can learn and the easier it is to connect information together.

Here's what learning about a new topic might look like for someone unfamiliar with the information:

What happens:

• Unsure how to integrate new piece of info.
• Adds a little, but not much to the understanding of the topic.

And here's how a master's knowledge structure evolves with a new piece of knowledge:

What happens:

• New information immediately hooks onto prior knowledge.
• Lots of connections to the new piece of information, making it much more memorable.
• Potentially makes connections between prior islands of understanding, leading to more 'eureka' moments.

So do we need to "memorise"?

Given our allergy to the word "memorise", as it evokes images of late nights learning flashcards off-by heart, I think it's important to define good memorisation and not-so-good memorisation techniques.

Not-so-good memorisation

Rote learning is one of the least effective methods of learning. It involves forcing your brain to remember something, usually based on tedious repetition. Rote learning attempts to force information into your brain that your brain simply doesn't want to learn. Particularly if the information is unimportant to you, your brain will always struggle to commit it to long-term memory and prefer to delete it to make space for more useful things.

Good memorisation

What's something you enjoy talking about? History? Politics? Music?

Now think of something you know about this topic. Did you have to learn that off-by-heart, or did your brain just commit it to memory without much effort?

Good memorisation involves a desire to understand. We need to tell our brain "We want to know this".

Good memorisation comes from understanding, not learning facts off by heart. In other words, by understanding the content, we commit that knowledge to memory much faster. What's more, that knowledge is much stickier now because we're not rote learning an isolated fact that has no connection to anything else. Understanding develops a web of information and relationships between information that results in a rich picture of the topic that the brain can navigate through.

But we should learn to solve problems, not memorise facts, right?

Problem solving is often mistaken as a process separate from knowledge. As if knowledge is the set of cooking ingredients and thinking is the act of cooking.

This naturally leads people to believe that if we taught everyone to cook, then it doesn't matter what ingredients they have or don't have. In other words, if we could teach people to be "critical thinkers", then we could do away with knowledge. You can look everything up nowadays anyway!

But this isn't how thinking works. Thinking and knowledge are not separate - they are both part of the same soup. Knowledge begets thinking and thinking begets knowledge.

It therefore follows that more knowledge = better thinkers.

So should we offload memorisation to ChatGPT and Google?

The answer is clear: Absolutely not. We need facts, figures, statistics and information in order to be able to think. If you offload this to any technology, you end up impairing your brain's thinking skills.

How do you know what to commit to memory? That's easy - whatever is important to you. Whether it's your work, your hobbies or just knowledge for the sake of knowledge, your brain needs information for it to think and develop understanding.

We'll see in a few years where this epidemic of "knowledge lookup" with ChatGPT leads to - I suspect nowhere good.