Roughly speaking, it represents the transition from an adder to a multiplier. A 1.0 programmer is competent at "adder" tasks (scripts, bug fixes, features) but not yet ready for infrastructural work. A 2.0 programmer is highly competent as a multiplier. It goes from 0.0 (complete beginner) to 3.0 (global multiplier) but most (probably 99%) professional software engineers are between 0.8 and 2.2, with a median around 1.1-1.2.
This isn't general enough to apply to computer scientists, nor do I intend it as an all-purpose ranking for software engineers. It's an approximate measure of a generalist's professional development. It's not really applicable to, say, machine learning experts.