The DSM developed at a time where different psychologists had different ideas about what causes mental illnesses and was specifically made to be agnotistic and not favor a specific school.

No quite...it's more about 'inter-rater reliability',

It doesn't even do that job well. They didn't run any studies for the DSM-V that investigated the inter-rater reliability of their new proposed categories.

At least on LessWrong, we could, perhaps, start using the research domains the NIH is proposing, instead of to-be-dated terms like 'depression' or 'schizophrenia' or 'anxiety' cause we're on the bleeding edge of things like that.

If you use a term like depression and use it correctly you can refer to a huge amount of existing papers on the subject that provide knowledge. Terms like autism to frequently get abused on LW.

Using labels because they are bleeding edge often leads to not really understanding the labels one uses.