As both a user of slicehost and google app engine, I wanted to chime in here. Google's approach for scaling is "do it our way, and you don't have to worry abou

As both a user of slicehost and google app engine, I wanted to chime in here. Google's approach for scaling is "do it our way, and you don't have to worry about architecture. You won't have to worry about master/slave nodes in your DB. Mirroring static files on separate app servers. Sharding your database by feature. Using a hardware load balancer router, etc." It is the instant coffee of scalable web apps. Just add hot water and enjoy. Slicehost, on the other hand, is the other end of the spectrum. You're roasting your own beans, grinding them up, boiling your own water, and so on. You get total control, but you have to know what you're doing and be comfortable with it. As an aside, its been pretty well documented that a small team of good developers can scale a web app very quickly, but Google's aiming more for the hobbyist market: Its trying to entice a single average developer to write a web app that will scale very quickly. And for the record, I don't see AppEngine as some kind of magic bullet either. But if I was a full-time CS student in my freshmen/sophomore year, and I was told (learn rails/apache/passenger/EC2/MYSQL/sharding/etc) vs. learn python and appEngine) I know which way I'd go. Its easy infrastructure with a low barriers to entry -- both in development and financial cost. (Lastly: Not trying to start a flame war over what's better. I'm simply trying to state a point that AppEngine may have more "hobbyist" appeal.)