,
11 tweets,
2 min read
Read on Twitter
1/ There's an excellent parallel between the idea of a personal "Second Brain," and event-sourcing in relational databases
2/ Relational databases traditionally had to be "architected" upfront, forcing you to think very hard about the data and which is the "correct" model, how the data should be connected, without real-world testing
3/ But now with event sourcing, you don't store the global model, which is huge and vulnerable to changes. You store only changes (called events or experiences) to the model
4/ By storing only the changes, you can construct a model that fits real data if you want, but if it no longer works down the line, just throw it away and use the changes to build a new model that fits the data better
5/ This allows you to embrace and even be hungry for more data, more events, more change, because changes don't threaten the global model. Your system is anti-fragile, it can handle rapid change
6/ For similar reasons, your digital second brain should only store the *excerpts* from sources that you consume, not entire source texts. Save only the parts that inspire or surprise you, because those feelings are signs of a "change" in your model
7/ Trad'l knowledge management was like trad'l relational databases, trying to build a comprehensive model of a person's or organization's knowledge base in a hierarchical, rigid format
8/ This never worked, not just because knowledge doesn't fit into a comprehensive taxonomy, but because as you approached completeness, your model got more rigid and resistant to change, slowing down your learning, making your knowledge obsolete
9/ So knowledge mgmt was inherently self-destructive. But PERSONAL knowledge mgmt puts control of each person's knowledge base in their own hands, and trains them to manage it informally using something like digital notes
10/ The informality is key: because you're free to store only fragments, snippets, and pieces of sources residing elsewhere, you have a record of changes that, like in event sourcing, you're free to fit into new models whenever needed
11/ This makes change no longer threatening – it is more events to fit into more models. Info overload becomes a thing of the past, because your sphere of responsibility is actually quite small: only unexpected changes
