You would think all the efforts towards social outreach would focus on poverty as their #1 criteria - the most direct and obvious criteria of people in need - but they almost never do. The problem is these programs are created by upper-class people in their media-social bubble, cobbled together from the arbitrary championings within that bubble. Unfortunately poverty doesn't have a strong, dedicated representative in these upper class circles, because only a tiny number of people born poor ever make it into the media or valley cocktail parties. Black, Hispanic, Women, Veterans - all have members both in needy classes and in upper classes to champion their own cause. But using this represented-in-the-bubble approach results in an arbitrary mix of random criteria that only half serves the purpose. There are certainly many Black/Hispanic/Women/Veterans in need and deserving of help, but there's also plenty who are trust fund babies who don't need it. Meanwhile, there are a lot of obvious groups left out of this list - Native Americans, poor Appalachian whites, etc. It seems ridiculous to explicitly help Blacks and Hispanics, but not Native Americans. We could go on and list a dozen more needy groups left out, but why bother when the correct criteria should just be poverty.
Discrimination on the basis of superficial factors exists. Positive discrimination on the basis of similar factors is inefficient for exactly the reasons you describe, but a) it's a lot cheaper and simpler than means-testing people to find out whether they're actually poor or just pretending to be and b) people often rely on superficial characteristics when choosing role models, which is natural because you want a model who's substantially ahead of you (otherwise you don't have much to aspire to) but the greater the gap you wish to close the lower the probability that you can establish personal contact with all the extra nuance and information involved.
You would think all the efforts towards social outreach would focus on poverty as their #1 criteria
This is a really, really, really terrible criteria to use. It is one with a proven track record of actively encouraging people to be failures in order to qualify for assistance and it becomes a trap they don't know how to escape. It is much, much, much more effective to define need or merit on some basis other than poverty. It is much better for society to say children, handicapped individuals, new parents, people recently fired...etc... deserve or need assistance and not "poor people."
I see your point about moral hazards, but in this case, I don't think anyone would avoid working in order to become poor in order to get free advice about how to work.
There are psychological costs to having people self-identify as being poor enough to qualify for a program that are not there for self-identifying as female or black or hispanic. You are asking people to say to themselves and the world "I am enough of a failure to qualify for help." It is not a basis for future success. It is the opposite. And it comes with all kinds of problems.
I wish YC approached the shortage of women and other minorities in tech/startups differently. But I don't wish they took the approach you are suggesting.
Following that logic, why is the tech industry targeting mostly white men creating harm? Why did the Boy Scouts banning gays create harm? Why did all-white country clubs create harm? Why did a ban on gay marriage create harm? We can easily justify any exclusive institution by ignoring the outsiders and weigh only the benefits to the insiders. Of course, this is painfully ironic when the purpose of the institution, in this case, is inclusion.
