Even in this down economy, finding technical talent can be extremely difficult. Résumés are worthless. It comes down to this: programming is a craft but we don't get to see the craftsmanship and creative process when we hire programmers. At least not before we narrow down the application pool of thousands to a short list of real candidates. After that, during the interview process we can do things like logic questions and programming exercises to see how the programmers solve problems.
The issue of finding talent is a very real concern not just for start-ups but also for established organizations that are selecting technology (like a CMS) that they must staff up to support. The natural tendency is to go by the numbers. Choose a technology with a "household name" like Java, .Net, or PHP where programmers are a dime a dozen. You are bound to find a needle in a haystack that big. Where this logic breaks down is that when a technology is so widely known and used, everyone seems to have it on their résumé. You can't tell if the candidate just spewed out a few hundred lines of crappy code or if they have immersed themselves in the technology to the point where they understand the core principles and philosophy. You can't tell if they are a talented craftsman and are able to learn new things. By selecting a household name (commodity) technology you have just made the haystack bigger without necessarily adding more needles. It is hard to find a Jamie Oliver by putting out a search for people who can make a hamburger.
When it comes right down to it, a language is just syntax that lets you do things like assign variables, create conditionals (if statements), and loop through lists. Learning what is available in libraries is the true learning curve when it comes to a technology and those libraries are constantly changing. So a good programmer is always learning. One of my favorite books about programming (The Pragmatic Programmer: From Journeyman to Master) recommends learning a new language every year to expose yourself to different approaches. The best programmers love the challenge of learning a new language or framework. They read books on different technologies and practices. If I was hiring a PHP developer, I may get an even better result if I found a great Java developer who is open to working in PHP. Of course, then I would have to find a great Java developer. If I was looking for a great Java developer, I could see myself looking for people who have worked in Python, Ruby, Groovy, or even Scala. That would show that they are committed to learning new things rather than churning out code the same old way.
When you look at it this way, the relative size of the "talent pool" should not drive you to a technology choice. You want to choose a technology that has proven itself to be more than a fad so that it will continue to evolve and mature. You want to work with a language that has good resources for learning (like articles, books, IRC, and mailing lists). But you don't want a technology just because a majority of technologists claim to know it. Most importantly, you want a good programmer. If a technology is sound and effective, a good programmer will be able to learn it and be productive with it.