Your misconception here is to think that valuation means how much the company is worth today. Or to read it as an intrinsic signal of the difficulty of creating what the company has created so far.
It's neither.
Early stage investors make calculated bets on teams that they think can succeed in the long run. Valuation is a technicality, and depends on how well founders can pitch their idea, and how desired they are by other competing VCs. Ultimately, what really matters, is the potential of the team, and the size of the space they're exploring.
In this case, Sequoia is not investing in a simple messaging app that can be "replicated in a couple of hours" (regardless if that's true or not). They're giving money to a team they believe can do something great with it. And one that spent 3 months being personally trained by Paulg plus some of the best minds in the startup ecosystem, to seize opportunities and create something big.
From a purely statistical perspective, 80% of investments will fail. It's very likely that Magic will fail, too. But the 20% that succeed will pay for all the other ones.
For anyone interested, I strongly recommend "The Launch Pad: Inside Y Combinator" [1], by Randall Stross. It shows in incredible detail the ins and outs of YC, and explains why angels and VCs do things that seem, at face value, crazy for outsiders. It will give you a totally different perspective on deals like this.
[1] http://smile.amazon.com/Launch-Pad-Inside-Combinator/dp/1591...