Sustaining an ecosystem of “truly copylefted” telematics applications through the “Work-for-hire” loophole of FLOSS licenses

Imagine Google decided to aggressively responsiblize and incentivize his employees, consultants and partner companies.
It would offer them very extensive performance bonuses on profits generated by software services, which they could design, develop and manage, by extending both publicly available FLOSS and unreleased Google-derivatived FLOSS.
It would offer them a “Consulting/Partnering Agreements” where it irrevocably stated to hire their services (as “work-for-hire”) to modify such source code; as well as to hold full responsibility for the managing and hosting, except for its obligation towards Google to respect a set of specific (mainly) hosting requirements. They would sign concurrently the assignment of all copyrights and other intellectual goods generated from their work on such code to Google.
Google, as the provider of partially-copyrighted source code, branding (and possibly some inevitable liabilities), would get 1-5% of the direct revenue generated by the service. The Partner, in turn, would get 95-99% of the revenue, sustain all consulting and hosting costs, as well as and fully “manage”, and be responsible for, the hosting of such service on behalf of Google.

It would also offer “Software Quality Review Agreements” to the same parties, in order to apply the same decentralization principles to quality and security assurance of its code. All such parties would be offered (under “work-for-hire” terms) a small symbolic amount to: obtain the code, review it and offer their feedback at their will.

Now, let’s imagine instead that all these would done by a foundation (or governmental entities, or a redundant network of national governments), which irrevocably commits to the following:
* any user and anyone in the world could decide at anytime to become a “Partner”
* all source code that is assigned to the foundation, and therefore that running on any of the partner-managed “derived services”, would always be available to anyone willing to sign (online or on paper) one of those 2 agreements.
* Requires that those Partners abide (possibly just above a certain number of “active users”) by severe hosting requirements (similar to the ones we have drafted here) for the running of those services, which concretely and enforceably places all hardware and software, running beyond the point of decryption, under the collective democratic control of its users.
* As the foundation reaches X thousands of active users of services hosted by it or by Partners; it would offer to each of those users to join other willing users in forming the sovereign body of the foundation, though a carefully designed constituent processes.

I am suggesting this may be a good thing, as it would create an ecosystem of developers and users where anyone, though partly limited from such democratically-controlled foundation, could access remotely-accessible software applications under “practically-copylefted” terms
They could start building new applications, as well as derivative works from any publicly-available FLOSS. Those works would never be “released” to anyone and, although still bound to the derivative terms of  GNU GPLv2 or other FLOSS license, would be made accessible to users and everyone under terms that amount practically to those of the current Affero, plus the means to collectively verify the code that is actually running.