On Learning Slowly
There’s a genre of content online that frames learning as an optimization problem. Learn React in 30 days. Master system design in a weekend. The implicit premise is that faster is always better.
I’ve been thinking about whether this is true.
Speed is sometimes the wrong variable # When I’ve tried to learn something fast — really compress it into a short window — I tend to end up with a shallow model that breaks under pressure. I can answer the interview question but I can’t debug the production incident.