It's a major shift I agree... but it's a great decision if executed correctly.
Every company has a set of main goals that imply that development can't be completely autonomous, but hiring good developers is not just about code it's also about people with a vision and a sense of purpose for your product. Take the way Github works for example: at the end of the day their main goal is an accessible hosted interface to git, to make git version control easier and ubiquitous; nevertheless having that main goal does not stop their developers from creating new features (say for example a desktop client) that make the product better though it may or may not directly translate to their main goal.
If you have a company that develops software and your developers have the autonomy to work on things that they actually believe will further increase the companies value (be it an interface change, bug fixes, or the refactoring of bad code) you will most definitely have a better result than if you make people work on things that might not be their forte or that are uninteresting to them.
Now I don't mean that people should have a complete free reign of the product, as some things are too important to forgo because no one wants to do them, but giving developers the chance to work on things they actually believe in as opposed to things that might not translate to creating value in their minds can go a long way towards good morale and retention, and un-doublty a better product because of it.
Every company has a set of main goals that imply that development can't be completely autonomous, but hiring good developers is not just about code it's also about people with a vision and a sense of purpose for your product. Take the way Github works for example: at the end of the day their main goal is an accessible hosted interface to git, to make git version control easier and ubiquitous; nevertheless having that main goal does not stop their developers from creating new features (say for example a desktop client) that make the product better though it may or may not directly translate to their main goal.
If you have a company that develops software and your developers have the autonomy to work on things that they actually believe will further increase the companies value (be it an interface change, bug fixes, or the refactoring of bad code) you will most definitely have a better result than if you make people work on things that might not be their forte or that are uninteresting to them.
Now I don't mean that people should have a complete free reign of the product, as some things are too important to forgo because no one wants to do them, but giving developers the chance to work on things they actually believe in as opposed to things that might not translate to creating value in their minds can go a long way towards good morale and retention, and un-doublty a better product because of it.