Excellent! I finally understand it. However, one question: at 3:40 where the TALENs are biding to the DNA, it seems that the TALENs are too bulky to fit close enough together to bind to successive nucleotides.
It is more specific than CRISPR, but all nuclease technologies have a degree of off-target activity. This introductory talk briefly mentioned obligate heterodimer mutations. These greatly reduce off-target cleavage arising from binding to only one half-site. With these FokI mutations, TALENs are arguably the most specific platform. Specificity isn’t the only issue though, new technologies like prime editing, which avoids creating a DSB, also reduce unwanted repair outcomes. See my talk on that...