I like to look at this two ways. A cure (which I still see as miles off and coming in two phases) and the automated management system, which I think the open source community will get to in about two years, with the commercial products not far behind. There's already a patent out for the combined sensor/cannula. The same company makes a tubeless pump which can be controlled via an iPhone app. Not too much work, then, to link that up with something like Loop, an AP app that can manage it.
My only question would be what you mean by 100% accurate for CGM. They're pretty close to that already, for the fluid that they read glucose in. What they aren't is a 100% analogue for blood glucose levels. That's the one bit of your crystal ball gazing that I think will see some of the largest changes. I don't think we're too far from CGM systems that are non-invasive. There's a lot of work going on there.
Faster insulin and Liver targeted insulins will speed up the time frame in which insulin acts, making auto sensing algorithms (which are already out there in the open source community) function effectively. The advent of longer life Glucagon will allow for dual chamber pumps. Who knows, they may even get Amylin into one of these and give us a full electronic beta cell stand in.
On the cure front, they'll work out how to stop the onset soon after diagnosis earliest. This will give respite to those who are diagnosed and they will be able to regenerate beta cells.
Those with longer term diabetes will also be administered the auto immune attack blocker. Some will recover some beta cell function. Some won't. Those who won't will need stem cell or implant therapy.
Just my two penneth!
