I was just thinking of a vision. A single person, sitting at Starbucks, sipping a latte or some other such drink. That person is intently looking at their laptop. They are, unbeknownst to any of the other patrons, starting a business that can scale to reasonable popularity and even take in money from day one right there at that table. The part of the vision that’s not common just yet is that there are 3 others doing the same thing at the other tables of that same Starbucks.

How are they doing it? With these:

• Amazon EC2
  • dynamic server imaging and bring-up for hosting your service
• Amazon S3
  • persistent on-line storage for backing up EC2 and serving content
• Google Apps
  • mail, calendar, contacts, IM, VoIP, word processing and spreadsheets for all your communication and collaboration needs

You need mail, IM, etc for a business and Google Apps has a compelling offering for startups. Right now they are the only ones with this kind of package, but they won’t be for long.

Assuming you can beg someone to give you their EC2 account (its still in limited beta right now), you’re good to deploy an app that can take a pounding from a Digg referral or the like.

You might still need some platform apps to really bang it out:

• Keynote
  • widely reported to rock PowerPoint’s existence and will help get that Series A round or tell your local user groups about how cool your stuff is
• Skype
  • some nice VoIP-to-phone integration features
• Firefox
  • no doubt

Obviously you’d still need your development tools and whatever software was required to build your startup’s software. However, most dev tools are free these days and lots of great software/libraries are, too. If you’re developing a software service, you’re most likely already in possession of most of the tools you’d use, anyway.

Assuming you were running for 4-6 months on the above software with one person, the approximate cost for the above for a Web app-based service would be about $2,000 - $4,000 USD, assuming you got somewhat popular in that time but barring explosive growth and assuming you aren’t dealing with gigantic storage requirements (e.g. YouTube).

The cost of trying something these days is basically a pretty short amount of loot. This really drives home what others have been saying, but I think it goes even further than that. Paul Graham is fond of saying that the best age to start a startup is when you’re young. I used to agree, but now I think that one of his main reasons is not applicable anymore. Any college professor could reasonably do it themselves over summer break if they were so motivated. Or a stay-at-home mom could easily build something over a 6-month period while the kids are taking their nap or on play dates. Hey, maybe Paul should start just sending people checks for $3000 after an initial vetting and letting them work from their own homes on their ideas for 3 months. Why should they have to move to Cambridge or San Francisco when any Starbucks, Tullys or Borders will do just as well?

With the new efficiencies coming out all the time for service hosting, business process streamlining and location-independent business, the playing field is getting level really fast. And this is also reducing the risk involved with doing a startup since building a proof-of-concept iteration can literally cost less than a new big-screen HDTV.