We've seen several threads on news.ycombinator on the need for a "founder's visa" and immigration reform. The current H1 visa system is cearly busted. If there is going to be a new office set up for early-stage entrepreneurs, this is certainly one of the important tasks at hand for the federal office.
http://news.ycombinator.com/item?id=815457
Eyeballing the whole thread, it looks like support ran 50/50. Note that, for instance, arguing for a "more Canadian-style immigration policy" is not the same thing as supporting the Founders Visa proposal. We are allowed to believe simultaneously that H1-B is busted and the Founders Visa is a bad idea.
Wait, no, here's a more recent thread about Founders Visas. It's got 89 votes. The top comment opposes founders visas:
http://news.ycombinator.com/item?id=832363
It refers to the HN thread on Graham's original post about Founders Visas. 196 votes on that post. Again, the top comment opposes the Visa:
Comment votes, top-comments or even prevalence of critical comments are not reliable indicators of net agreement/disagreement. Threads have lots of factors that cause criticism and disagreement to be over-represented, or cause quibbles to be misinterpreted as opposition.
I myself would criticize the 'investors panel' mechanism because I support the overall gist of any policy to increase founder immigration. Sure, I understand bokonist's critique and I'd prefer saner non-political mechanisms for boosting beneficial immigration.
But given that immigration policy will continue to be a bureaucratic mess -- there's no 'grand fix' proposal on the table as an alternative -- even a flawed program could be worthwhile. All the existing 'doors' through which immigrants enter are politically-flawed; why not have one more door, also admittedly flawed, through which a few more valuable immigrants might enter?
The assertion that creating a founders visa enables more successful startups that in turn makes the economy grow (preferably sustainable growth) has not been proven at all. Its a very nice assertion, and I'd like to buy into it, but I have seen no meaningful data in support.