The concept Dickson Despommier sets forth here is a really cool one:
BUT, and you knew there was going to be a but because IFs and BUTTs and such, there are a multitude of factors to consider in farming in this manner and in producing food in general. As the clip is just a few minutes long, I'm pretty sure Mr Despommier has taken into consideration most of these things and all that's left is making sure that when building these sorts of buildings, the usual shortcuts aren't taken. Which brings me to the first point: A structure that produces both food and electricity in a manner of a small microcosmos (narf) has to take into consideration the structural differences those areas require. Fish farming requires large areas of water that can't leak into the housing area or directly into drinking water or electrical wiring. In high risk earth quake areas? This would be an even larger factor.
Growing other types of farm/edible animals and hygiene in ventilation as well as breeding areas is a concern. Even if you're recycling the waste material inside the building, the recycling has to take into consideration the same sort of safety issues as the water tanks. Will the animals be slaughtered in-house or will livestock be limited to animals who produce edibles but are not so much eaten themselves? Again: recycling, hygiene, cross-contamination. I think it's safe to say at this point that this type of building will demand changes in building codes in a lot of places where the use of a space is strictly regulated.
Moving on to the plants. Again structural issues become the focal point. When building something like this, you absolutely cannot cut corners. With larger greenhouses indoors, humidity and molds in other floors need to be considered. Appropriate ventilation and water blocks and none of the type of building that's been on the rise in some places (not naming names) where a building is erected and needs to be thoroughly redone within a few years because of material deficiencies and unreasonable schedules during construction. Similarly with the power production. Alternative methods of using things like the heat animals/plants/people produce on their own should be explored.
And lastly, the most esoteric of issues: what will be the quality of the food produced by this small ecosystem? A lot has been said over the years about how our soil is getting depleted of nutrients due to intensive farming and the use of lab created chemicals over more natural methods of pest control and soil enhancement. Whether that is in any way true or not, I couldn't honestly say. There's data on all sides proving all kinds of things. However, you do have to question how well animals and plants would fare in a totally artificial environment. Well, besides fungi. Apparently they do splendidly pretty much anywhere. How sustainable and nutritious are crops farmed on a soil half a meter deep, devoid of elements, seasons and a natural source of sunlight?
Self-sufficiensy is or as near as, perhaps obviously, the way to evenly distribute the world's resources. The Problem has almost never been that there wouldn't be enough food to feed everyone, but that the food, like all resources, is unevenly distributed because of artificially created needs. Needs that are NOT food related. A simple We Should All Go Vegan is not going to resolve the fact that individual bodychemistry makes it impossible for some people to thrive on veganism, or the fact that there are a great number of areas in the world where plant food is simply not an option all year round. Would the world be able to house AND sustain enough farm land to house 7 billion people if we all went vegan and just lived in the areas of the world where that's possible?
So, I guess we need to look at building stuff like this and consider the structures that support those needs that maintain the inequal distribution of wealth.