Do the PhD and work out your AML software company idea while you are doing the PhD. You can drop out when you're ready to run with your startup. I personally know about 8 people who have become software millionaires this way. It's also how places like Google and Yahoo! got started.
The "Real World" is totally overrated. Entry level software jobs generally suck donkey balls, even if they sound like they won't.
However, it is very particular to your situation. If your job offer is to work on Photosynth at Microsoft Research, turn down the full ride to Toledo State U. However, if you're funded at Stanford or MIT, turn down the entry level code monkey job at Amazon or Yahoo!
Thats what I am doing except I am doing research right now on computation as an undergrad and I am auditing classes to get the information to complete the research. Its fun when you can pass out of classes to simplify your course load (self study).
The "Real World" is totally overrated. Entry level software jobs generally suck donkey balls, even if they sound like they won't.
However, it is very particular to your situation. If your job offer is to work on Photosynth at Microsoft Research, turn down the full ride to Toledo State U. However, if you're funded at Stanford or MIT, turn down the entry level code monkey job at Amazon or Yahoo!