Horses for Courses, I guess.
Seems that what potential employees value from Google and Microsoft are different.
Update : read this.
Here's my take :
We're back to John Hagel's unbundling of corporations into infrastructure providers, customer relations and product innovation. Google still doesn't quite know what it wants to be. Its success is based on big infrastructure, but its culture is one of product innovation.
An open internal market encourages that ... but there are some people who want to work in infrastructure or customer services, they might be happier going back to Microsoft. OTOH, as they go to M$ they'll help make it less innovative.
Google could find itself wrenched between two imperatives ... if it goes infrastructure, it will see the innovators slowly evaporate. If it holds on to the innovators it will see service reliability decrease ... allowing attack from elsewhere.
One solution might be to partition itself into different kinds of companies with different cultures. Another is to split entirely. Google outsourcing infrastructure to Microsoft would be ... well, a turnaround. (Not that I'm predicting it.)
M$, also need to decide, of course.