> Python 2 will never die.

Everything dies.

> Ask Kenneth Reitz for example. He says he will use Python 2 forever.

So, assuming this currently-stated intent never changes and Reitz is immortal, Python 2 will always be in use.

The first of those premises is somewhat suspect, the second even moreso.

> There are a lot of companies with huge codebases who never will port to Python 3.

That's probably true. Its less likely true that those companies will never port off Python 2.

> Everything dies

Python itself might die when Python 2 dies, and arguably, Python 3 makes that more likely.

In my job we have some big projects on Python 2 and Cython gives us access to some of the "new Py3 features": types (real types, not just hints) and async IO (through C primitives). We eventually chose Cython because at least we can get big speedups (depending on the code and how much can be properly cythonized, of course). With Py3 it was a change to the codebase (which we also had with Cython, of course), to get approximately the same performance, maybe slightly less, maybe slightly more.

This is the unfortunate truth. My company still uses Python 2 solely, not a single line of Python 3.

Unless there is some major breakthrough in porting tools (in combination with more and more libraries supporting 3), Python 2 will still probably be the only Python version in many enterprises for the next 10 years.