I learned SQL around 1996 or 1997 and had an absolutely fantastic college course or two involving it. I've mostly been a backend developer... having the abilit

I learned SQL around 1996 or 1997 and had an absolutely fantastic college course or two involving it. I've mostly been a backend developer... having the ability to go in and fix SQL and make things run in < 1 second instead of 5-10 minutes has been one of the best skills I could have picked up. It is sad that so often "self described 10x programmers" build solutions to go around SQL that are horrible failures. Poor use of ORMs, weird abstractions at the application layer that force developers to use poor data access patterns, unnecessary "locking" at the application layer, unnecessary "existence checks" at the application layer, processing objects 1-by-1 in the application. All these things lead to terrible performance and huge wastes of application memory/IO. I love some of the NoSQL solutions too as in some cases they can force teams to use better data organizations/patterns and scale so well. Those patterns are often possible in an RDBMS but the system doesn't guide a team to using those patterns. The way CQL in Cassandra forces you to think about data organization is great for example.