App development is something you can pick and contribute to fairly quickly. Similar to say building Facebook.
Contrast that with 'real problems'.
You want to cure cancer, build a new generation of graphics technology... then you need a whole swath of domain level knowledge that many companies today don't train for. More than that, when that technology is done, you can be easily discarded. So what's the purpose of getting deep into a really deep technical domain?
As to why not fix healthcare.gov? I'm pretty sure even the youngest and most naive tech student knows to stay away from CGI or Accenture... Sadly, these organization know how to get government and enterprise contracts. You're not solving any real problems with healthcare.gov. You're playing bureaucracy and checklists and billable hours.
You want people in the 'real problems'.
Train them, ensure they can have a long term viable career in that field, and pay them decently.
It's not rocket science.