Let's talk about Curation

Curating Books in a Library

In a previous newsletter that I’ve written, I explained that to survive in the 21st Century, with the advent of Generative AI and the content explosion that it brought about, we need to do Critical Thinking and Curation (Two Cs).

I wanted to get into Curation a bit more after having used LLMs for a bit now for my work and research.

We all know LLMs can hallucinate however its ability to generate largely relevant and large amount text, and within these texts are ideas and topics that may take you a long while to gather really tip the scale towards the benefits, moving us to think about solutions to manage its hallucination rather.

Here is what I found out though. While it can generate a lot of ideas, perspectives, topics etc, each one of them will take time to ascertain further whether it is true, relevant or to summarise it in a more succinct way, we can use it for our cause and reduce the risk of bad consequences.

Curation takes effort. While LLMs help with the gathering, researching and extreme typing speed, reduction in research, idea and text generation time period, but it may increase the time to curate each idea. To give a simple example, for instance you might spend 3 to 5 hours for research, generating 10 ideas and turning them into text notes. Now, using LLMs you can cover a larger base, getting 15 ideas perhaps. Now assuming each idea takes about 0.5 hours to curate, that will mean about 7.5 hours used up for curation even though you do not need to turn them into notes.

While time is one factor, curation also requires us to exercise our brain. If you have read the book, “Exercised” by Daniel Lieberman, you will know that our body and even our brain which consumes a large amount of our energy intake are not built to exercise so a very mindful effort is needed to start and maintain the momentum of thinking.

Ok now what? Does that mean that LLMs are bad then? Not really. Again, the pros of LLMs, at its current state, definitely outweigh the cons of using it. So the question I moved on to was, “Can we reduce the time for curation then?” One word came to my mind, “FAMILIARITY”.

The more familiar you are with a certain topic, the better you are at curating the particular topics that are generated. Once you are familiar with a topic, you develop an instinct to determine if a topic/idea is feasible for your cause very quickly.

This is how I see it. We cannot be experts in everything, that is not going to work well. However, we should find out who are the experts in the area we are researching on. We should ensure that these experts have the relevant background and experience for us to trust their own “curation”. With that, we then use search engines to look for content from these experts to assist us in the curation, for instance if the idea/topic is commented by the experts we can quickly take it in.

What are your thoughts on this? Feel free to comment in the comments.