It may be more likely than I think, but what matters is if it is so much more likely that P(Silicon Valley startup getting funding) < P(Canada/foreign startup getting funding), all other things being equal. Do you think that this is the case?
I don't think that every startup needs to be in the Valley to get funding, but what I've witnessed here is a prime example of "it's not (only) what you know, it's who you know". If you have a viable startup, knowing people that have made names for themselves in the startup scene, many of which are based in Silicon Valley, can go a long way towards getting introduced to potential investors, whether they be angel investors or VCs.
The idea is a good one, but I'm wrestling with the dollar amount that needs to be invested before the founder can get a visa. Specifically, how likely is it that a US venture capitalist will invest $250,000 or more in a startup that's not located in the US (or Silicon Valley for that matter) vs. another startup that is?
Even if the non-US startup will move to the US once they receive funding, the largely intangible advantages bestowed on a startup in the Valley (which I've witnessed in the month that I've lived here) give it a better chance.
To be successful in facilitating the immigration of founders to the US, the founders' Visa should be available to any startup that has received any funding from a "qualified US investor" over $15K, which would include non-US startups accepted into YC.
And to avoid creating a huge immigration loophole (which seems to be the author's rationale for suggesting a "real money" cutoff), perhaps the Visa could have a clause that limits it to two years unless additional capital is secured and the startup is deemed to have "passed muster", at which point a permanent resident visa could be granted.
I think it may have more to do with the latter. A quick Google Scholar query of "why is computer programming dominated by males" provides some interesting results, including an abstract with the key finding that "men had more confidence in using computers than did women even when statistically controlling quantitative ability. In fact, female CS majors had less computer confidence than did male non-majors." (Beyer et al., 2003)