I think David Blume is overly fond of hyperbole, but some of his numbers do hold up. I do believe that 3-10 pounds of food per square foot is a reasonable goal, for example. certainly not in all conditions or in all climates or without a lot of thoughtful design, but my experience leads me to believe that he's not entirely out of line.
just a quick example: the small farm I work for grows some leeks. I choose leeks because we space them very uniformly. this is in Fall
City, WA, about the same latitude as Seattle. including paths, each leek has a little under one square foot of dirt to grow in. by harvest time, they're generally about a pound each after cleaning. that's roughly one pound of food per square foot. we till the dickens out of that dirt and keep it bare between the leeks. fertility is addressed with
chicken manure. vetch and rye is grown in the cool season and tilled in. in other words, substantial improvements over that 1 lb/ft[sup]2[/sup] yield could be had without much effort. put some real thought and observation into it, and 3-10 lbs/ft[sup]2[/sup] doesn't seem far-fetched to me. give it a little more effort and that 3-10 pounds might include enough balanced nutrition to actually nourish folks.
I have never measured yield/area for my own projects because it's so much more complicated to do with things all mixed up. the Dervaes crowd does a good job of that, I suppose, and it looks like they're getting a little over 1.5 lbs/ft[sup]2[/sup]. my impression is that the error Blume makes isn't to overstate potential yield, it's to exaggerate how easy and common it is to obtain that high yield. 3 lbs/ft[sup]2[/sup] might be relatively "easy to come by" as Blume states, but 10 lbs/ft[sup]2[/sup] probably isn't in temperate regions.
the tropics are a whole other story, as Fred Morgan made clear. one jackfruit tree growing alone would easily put us well within that 3-10 lbs/ft[sup]2[/sup] range without any further consideration.