I started with Python 1.4 but the 1.5.x series came out shortly so most of my early work was done on 1.5.2.
Some historical comments, with probably a few errors....
When 2.0 was released, it included support for Unicode as a separate data type. 2.1 and 2.2 included many other new features but also started the discussions about maintaining backwards compatibility. These discussions led to the introduction of "from __future__ import <blah>" as the method to introduce new features that could introduce incompatibilities. In some cases, the new feature became the default in the next release. "as" became a reserved word around the 2.3/2.4 timeframe. It was first introduced via the __future__ mechanism and then became a reserved word. (And yes, the change broke one of my programs.)
The change in behavior of int/int was (somewhat) controversial. The risks for quietly changing a program were considered to high so the default behavior was never changed in the 2.x series. The also started the concept of a Python 3.0 release that could break the backwards compatibility guarantee.
I now use Python 2.7, 3.3, 3.4, 3.5 and 3.6 for testing. Any serious programs are written for 3.x
p.s. I still have a copy of the "Internet Programming in Python" book that was co-authored by Guido van Rossum.