I generally agree with the sentiment. I think it's very easy to each year just look at the standard prize amounts and decide that slightly beating the standard can't hurt, but over time that leads to higher and higher prizes. Certainly $1 million is absurd. We (PennApps) are at $10k for 1st place, which I think you could argue is high (keep in mind that turns into $2.5k/person). But we made a conscious decision not to raise our prize any more than that about a year ago, so at least it's not getting worse on our end.<p>I'm not convinced that giving already released hardware as the main prizes is a good idea. From my point of view, people don't choose to compete for the main prizes like they do for sponsor ones, and forcing them to take an Oculus Rift or something of the like that they could just buy if they wanted seems a little suboptimal. However, getting unreleased hardware is a value add in my book, and was something we considered for the latest PennApps with Myo, although they couldn't make the delivery date in the end.<p>Sponsor prizes are tricky. One thing we've done at PennApps is encourage more vaguely themed prizes rather than ones that reward the use of a specific company's API. OTOH, we're not ready to entirely restrict API-centric prizes since I do believe that for less "sexy" companies it is a value add (the writers behind this article both came from arguably "sexy" companies), and the competition for sponsor dollars is such that the negatives mentioned in the article about sponsor prizes do not outweigh the cost of removing them for us.