ETL vs ELT: what is better, and why?

In Episode 4 of Data Pipeline, Ilya breaks down the difference between ETL (extract-transform-load) and ELT (extract-load-transform): with ETL you transform the data first and store only the results, while with ELT you store everything raw and make sense of it later. Which is better? As he puts it, it depends on the use case and on factors like your storage space and the complexity of the processing you need. This episode pairs with the LinkedIn post that explains the same idea with an apples-and-juice metaphor.