OK, it's faster, but I don't care about speed, I care about accuracy. Last time I had something sequenced we were constrained by the 500 bp limit in the accuracy of results and had to(for confidence sake) sequence our sample in 300 bp blocks. We'd get the results in a day or two, but the small lengths we were dealing with made it prohibitively expensive (because of primers mostly). If this technique could do longer strands, it would be more valuable.

It does massively parallel sequencing. So for example, you could sequence your 500 bp fragment 100k times for ~100% accuracy. But large scale sequencing is where this would be valuable.

For example, imagine doing directed evolution of a microbe and then sequencing to determine the genotypic effect.