(2) 6000 dry lbs/acre/year = 3 dry tons/acre/year which is an extremely low yield. Even miscanthus and switchgrass get over 10 dry tons/acre/year, energy cane gets to 20 dry tons/acre/year and our grass gets to 25+ dry tons/acre/year. So that brings your $100/acre/year up to $800+/acre/year. Then for the chemistry it's important to note that much of the hydrogen gas produced is actually coming from H2O that reacts with carbon in the cellulose to produce 2 H2 + CO2. So, stoichiometrically you get significantly more than the elemental hydrogen content of the grass itself. That gets you another factor of 2 or so... and then we're at the $1750/acre/year mentioned in the parent comment.
(3) Agreed the transportation costs are better for ammonia, but we aren't actually transporting the hydrogen except over a feeder pipe into a refinery or ammonia plant. It's cheaper and simpler to transport the grass as opposed to the hydrogen, mostly because you get to avoid the pre-transport compression energy and losses. Again, as in (1) the issue with ammonia is the heavy capex based around Haber-Bosch pressure vessels and compressors... we didn't have any good ideas for reducing those costs, so there's no sense in competing there.
(4) We weren't at YC's demo day... not sure what you're referring to ¯\_(ツ)_/¯