There is literally nothing more powerful than having someone within the business introducing you.
Your points about specialising are 100% valid and it is very good advice however I would add that people need to think carefully about what they want to specialise in as if you make the wrong choice you are stuck with a career you hate and will struggling to move out of that speciality.
It occurs to me that this may be precisely the reason why specialists command higher rates and find it easier to sell their services: Specialization is a risk. Sometimes it's a big risk. The reason why you pay extra for a specialist in $FIELD_X is that you're compensating that expert for having taken the risk of specializing in $FIELD_X.
There were NeXTSTEP programming experts in 1998. Those people sure ended up hitting the jackpot, though nobody could have predicted it at the time. I assure you that there were also AOLServer Tcl programming experts in 1998. Nobody remembers them. :] (At least, not for that.)
This line of thinking makes me wonder if the standard "be a specialist" advice is yet another example of sample bias: Because the lucky and successful specialists make money, stay in business, and eventually retire and spend their time writing articles and talks about how great it is to be a specialist, we may think specialization is a better idea than it actually is. Make sure to balance your approach by talking to some of the poor folks who specialized in Pascal programming, or Palm WebOS. Or ask an expert Flash programmer how they feel about the rise of the iPad. Keep your escape routes handy!
You can get trapped by a specialty, but there are ways to get out of that. Open source opens up a lot of doors there.
For two of my three post-college jobs I had no professional experience in the area that I was going into, but had a decent body of open source work (on projects that were being used already by my future employers) that established that I was a credible specialist.
If you're unemployed and looking for a job, it's not a huge gamble to roll the dice and try to learn a lot about an area you're interested in. Just do so realizing that you may have to repeat that process again at some point.
As a side-note, just because a specialization is seemingly on the way out doesn't mean that it's not lucrative. I remember when in college, in 1999, I was probably one of the youngest people in the world that knew COBOL. The offers that I was getting at the ripe age of 19 were $200k+ because everyone was terrified of Y2K stuff. Since then I've mostly worked as a C++ developer. While there's less new software being written in C++ today relative to a decade ago, it's still a very marketable skill precisely because the number of people that are good at it is smaller than the number of jobs available.
Be careful specializing on technologies; that's a hamster wheel of a career. Better to specialize in application domains. Lord knows where ObjC will be 10 years from now, but search engine discoverability is probably going to be important for a long time.
This has gotten my 90% of my fulltime jobs, and 100% of my contract gigs.
If you are well connected, job boards and listings become a distraction.
If you are not, there is still hope. This is a highly valuable and transferable skill (and part of the overall career-building skillset involving self-marketing and interview skills).
Being well-connected isn't just for the rich and lazy, it helps in all aspects of your career.
Your points about specialising are 100% valid and it is very good advice however I would add that people need to think carefully about what they want to specialise in as if you make the wrong choice you are stuck with a career you hate and will struggling to move out of that speciality.