I agree.
If you follow the second order links down to what Disney actually did, they outsourced their IT to a contracting agency.
When they did this, they laid off 125 full time employees in the process, and between three of the contracting agencies providing the services to replace them, there were apparent;y 65 H1-B applications in the last 3 years. Presumably, not all 65 went to Disney, because the contracting agencies contract services out to companies other than Disney. In fact, a vast number of dark data center porn and shopping sites are located in that area of the country, down by Los Angeles, where the majority of that kind of content is produced.
What this story is actually about, is complaining that the full time workers were replaced with contractors, some of whom were probably in the U.S. working for the contracting agencies on either H1 or L1 visas.
The summary is a gross misrepresentation of the facts here, and going with a contracting agency is a valid mechanism for ensuring "Just In Time" capability, without over-employing in order to handle upsurges in workloads. It's how janitorial and security services are handled (when you have a large company event, you have the contracted agencies put on more security people for the event itself, and added janitorial people post-event to clean up afterward.
That said, the usual route a decent company will follow when out-sourcing to a local agency, as opposed to off-shoring the work entirely, is to require that the contracting agency hire a certain percentage of the workers that are being laid off to replace them with contractors. This has the effect of ensuring continuity of service, providing a built-in mentoring capability to the contracting agency for the processes and procedures being contracted out, and in general providing continuity of employment for at least some of their existing staff.
It falls under the category of "Not Being Dickish About Switching Over To Contractors".
But the idea that they should not be switching over to contractors at all, for something like IT services, which are generally modular, replicable, and have uniformly applicable skill sets, if what you are spending your time doing is pulling wires, spinning up VMs, installing system software on replacement desktop/laptop machines, and so on, is patently absurd. These are "cog jobs", where any sufficiently skilled cog can replace any other sufficiently skilled cog in the machine, and you probably won't lose a marching step over the replacement.
That, and surge scalability, make them rather ideal for out-sourcing.
Frankly, I'm surprised companies like RackSpace are renting out their IT people, rather than forcing everyone to live on RackSpace racks; it's a pretty ideal scenario for them, in terms of profit per employee, and gives them buffer for their own internal surge scalability issues. They get borrowable capacity, and other people pay to maintain that capacity at a certain level.
Add the fact that a lot of deployment is on OpenStack with standard deployment tools, no matter if you're working on your cloud or working on someone else's cloud: all the tools are the same, so all the skills are pretty much transferrable.
This is kind of what happens when you sufficiently commoditize an industry through standardization.