Why limit the number of labels

In the data modeling course in GraphAcademy, it is said:
« You should use labels wisely in your data model. They should be used if it will help with most of your use cases. A best practice is to limit the number of labels for a node to 4.»

Can you expert provide details about why it would be detrimental to have more labels?

Is it for memory consideration as the chatbot seems to propose? If so, is it due to indexes?
Is it for human readability?
Is it because having many labels suggest that our data should be modeled differently?

Hello. Neo4j allocates 4 spaces for labels in its node store. Once you go past that, it starts allocating extra space, chewing up memory and slowing down performance. You can read more about the specifics and recommendations in this blog post. Hope this helps!