We also discuss current bottlenecks for mammalian genome writing (mainly throughput and size), mechanisms to address them, and future directions for the technology. We think it is particularly poised to revolutionize generating training data for genomic AI models.
