That would fit and work on vehicles though, which would let them drive at night without visible lights. So still some good military applications, which means it'll probably still get a lot of funding and someone will probably manage to shrink the laser down quite a bit more. At least with a backpack-able version anyway.
Military vehicles can already drive at night without visible light. Except in the most extreme case (new moon and heavy cloud cover) - and then they can illuminate using infrared. Such a device costs <2,000 USD and is already wearable. More expensive and/or vehicle-mounted systems increase that capability to the point where you basically never even need the IR illumination, unless the vehicle is fighting inside a building (rare).
The technology in the article, aside from being heavily editorialized, will remain inferior to that for a long time. However, one possible application would be the use of higher-wavelength infrared for the active illumination, so that other militaries with night vision are unable to see your infrared headlights and then blow you up.
Mike Koeris Director, DARPA BTO YC W22
Not to mention that the requirements in the solicitation for speed, accuracy, and length of product are each at least an order of magnitude above what is possible in current in vitro oligo synthesis. And that's just for the intermediate, 19-month goal, much less the later ones.
Certainly, DNA synthesis has been a limiting factor in bio R&D and both the cost and turnaround time have remained fairly stagnant, especially for gene-length products and given the near-term explosion in demand from AI-designed proteins.
I can appreciate that DARPA is a fan of moonshots but would it be advisable to break this into sub-projects such as generally improving DNA+gene synthesis, developing new methods for cell transfection, error correction, in vivo assembly, etc? Focusing on in vivo and light-directed only (and together) might not be the best path forward, though it certainly sounds sci-fi.