Think small, don’t over-think it.
Go with the PC and use the Unreal Developer Kit.
Start working, beginning with the game mechanic, not the setting or story.
Learn from the mistakes of games that failed.
Have a business plan. Get a biz guy!
Learn about the psychology of games and soak up the evil of stickiness.
Toss away ideas that aren’t working.
Be objective.
Once you have your idea in place, staff up slightly.
Don’t extend the project’s scope.
Don’t solve problems with feature creep.
Are you getting along with your team? Do you have any douchebags?
Build, test, repeat.
Self-promote your game.
Tweet well known developers, like Notch or myself - if we like it we’ll retweet it.
Go to events like PAX and get into IGF.
Be willing to start over and learn from mistakes.
Don’t overscope.
Save your money when you get it, because it’s a volatile market.
Remember Rovio made 51 games before Angry Birds.
Your early games will suck.