There are known workarounds that make diabetes-1 not fatal; "doing some hacking" (whatever you'd mean by that) would likely have a significant percentage risk of being fatal if you want do do anything that does something to your body at all without spending, say, less than a year.

"Above all, do no harm" makes for a very conservative approach of being really sure before risking actual action.

They do "hacking" etc in mouse models but it is expensive and time consuming. You might have about 700,000,000 genomic variants to choose from so we need to do other studies to try things.

For the time being one of the most effective means of studying a disease is to sequence a pedigree of individuals some of whom have the disease and analyze the inheritance patterns until you narrow it down to the point you can be sure.

You can also do studies where you just sequence a ton of people with and without the disease but those can be more problematic statistically because of the number of possible variants.