"My 2 cents about why women are underrepresented in software industry is that not too many of them are interested in the beginning.
One big drive of male programmers, according to my observation (hundreds of samples), are video games, myself included."
I think you are drawing a false statistical conclusion here.
The basic hypothesis you are making is: "In the population of women, there is a lower average aptitude for programming than in the population of males". "This hypothesis is based on sampling my coworkers, and the general population".
Let's take as given that the statistical aptitude for programming in the general population is much lower than in the population of professional software engineers. Basically, if all of the women you've met and are using as a basis for your hypothesis are not your co-workers, then you are sampling from completely different sets.
If we presume you are familiar with n women (and they are picked randomly from the general population) then yes, it is very likely that they have a lower aptitude for programming than the set of your male colleagues. But this lower aptitude would apply to any random set picked from the general population irrespective of sex, or any other factor.
" it is the early stage education system to introduce programming to wider range of people"
Yes, this.