Frankly? It's a good study with a fairly sober assessment of difficulties, if imho overeager in extrapolation of future prospect. You can only do so much for buffering electrolytic medium without impacting purity (and therefore efficiency), and -- at those scales -- their use of sodium and potassium as reagents will only get costlier. They also incorporate tin nanoparticle as a catalyst, and even though it's by definition recoverable, that thing goes for $2k a kilo for shitty 100nm stuff and needs to be cleaned between batches. I could go on.
But that's the game: fluff it up for funding, hope it's cost-effective enough for someone to pick up. My real beef is that it's from MIT, a place that has a world-shattering breakthrough every other week for technologies that are hailed as success worth following up on a year later if they almost make it past the proof of concept stage. Seriously, if 10% of those announcements paid off, I'd be exploring Oort cloud in my fusion-powered spaceship made from superconducting paper-mache and plastic waste. And it's hard to not be jaded about it.
I want to be wrong so fucking much, though.