Eh? It's early days, and he's approaching it more like an engineer than a scientist; but it looks like he has the right end of the stick (for some sort of stick). He's synthesizing his food de novo, so it's very easy to control the inputs and eliminate confounding factors. He's also doing occasional measurements (albeit not regularly enough for Proper Science, perhaps), and he is offering his replacement for free to other people on the condition that they do the same measurements. His hypotheses are not ideal, because he's not a domain expert yet, but that just means he's going to be a bit slower until he reads more papers and starts avoiding other people's mistakes ;-)

One key thing I learned in university: Once you start reading papers, you're quickly going to turn into the world's leading expert on the tiny corner of science that you're reading on. (Of course, this has nothing to do with being a genius, and everything to do with no one else ever having cared about that corner before. :-P )