There are a few ways this project could be really cool/useful. As discussed previously this project likely couldn't compete with traditional synthesis technology.
DNA or RNA prototyping. For short sequences, like for small PCB designs, it's a pain to send it to an external fab to have it professionally done. It's often easier to build the more error-prone prototype by hand. Once verified or made in higher volume, then it's sent off to the fabs. Local, low-volume fabs always have a place in prototyping. Even now I still buy radio shack copper protoboards for one-offs or when only making 5 boards of something. Synthesis of short sequences is a similar idea.
micro RNA. Already short sequence synthesis. Perhaps with many combinations. So rather than fabbing 1 sequence with volume of 1 billion, I want ten thousand strands with volume 100, and error-prone fabrication is maybe OK. Local cheapie fab is also a good fit for this, compared to precise high volume remote fab with longer turn around time.