I still believe multi-threading has its value, and I don't think it's wise dismissing it.
The main problem scripting languages like Python have is that their standard libraries are written in C ... this is a technical debt and always the main culprit for the slow evolution of their VMs. As far as Python is concerned, many people aren't using Jython/IronPython/PyPy because NumPy won't run on those.
I believe the future lies in projects like Rubinius / PyPy ... but it will be an uphill struggle ... it's easier to get rid of scabies than to get rid of that GIL. And they've missed that opportunity with Python 3.
And from my reading he's saying "(hopefully) we don't need no stinking multi-threading cause multiprocessing will be how it's done".