Personally I like leading commas for the ergonomics rather than the aesthetics. When I'm developing or debugging a query the first column is typically the one I'm least likely to change. I tend to build up the query from there, so the last columns are the ones I'm most likely to change or to comment out. Plus I find it easiest to interpret the result set when columns that I'm using as a temporary reference are at the very end of the row. So for the way I work, I've found that with leading commas I don't have to do as much futzing about with commas.
My job involves a lot of ELT pipelines though and the queries I'm writing are often to transform a client's data from whatever ill-conceived data model they've been using to a standard data model that we use for all clients. Those queries require a lot more "detective work" to get right than the queries that run against our standard data model. If I was just writing queries against the standard, I'm not sure I'd spend enough time developing/debugging to really notice any ergonomic benefit.
