He does make a point but hes arguments to back that point are flawed. The reward<->risk exchange is not what makes mining necessary. Its the decentralized and TRUSTLESSNESS of bitcoin that makes mining necessary ... and (very) expensive.
For me BitShares, in contrast to Bitcoin, increases efficiency, due to no mining, but theoretically reduces the trustlessness of the network because you predefine (elect) those that construct blocks. Hence you cannot participate in the block production but just setting up a machine and doe some stuff (e.g. mining).
However(!!!), Andreas fails to see that the trustlessness in bitcoin is gone for years already because solo-mining is useless and people have to pool together their hardware constructing guilds of miners (i.e. pools). These pools are, by the very definition, centralized and can be controlled (at least in theory) by a single entity. And what is even worse is that with the mining-reward and the "luck" required to find a block, profitable businesses around mining are forced to join the biggest pools(!!!) to make the biggest profit from their electricity. This inevitably leads to having just "a few" pools.
What really disappoints me is that Andreas tells the MIT students "whenever someone tells you .... they don't understand"! For me this is a absolute no-go. Either you have the arguments or you don't. But stating that someone else does not understand just because they dislike (or only disagree) on this is plain wrong. Sorry to say this Andreas, but for me you became a puppet in the very moment you tried to influence these students. I hope they are intelligent enough to find their own opinion and do their research.