With what model are you comparing them? Lasers are not energy efficient with respect to how much power they output compared to how much you put in, but that is also not how you measure the efficiency of a pesticide. What you want to know is the energy spent per treated m2 of field, and I don't see how you can immediately dismiss targeted lasers over uniformly sprayed pesticide without further analysis of the average number of weeds in an m2 of field combined with the power requirement to kill it with lasers.
Lasers are going to use far less energy than pesticides per plant killed.
A 20W laser (120W input) can cut 3/4th inch thick pine at 6 inches a minute. For burning through a young weeds stem you’re talking fractions of a second per plant. It’s likely the motor and other electronics ends up using more energy than the laser.
Mechanical methods are going to be a lot more complex and thus have significantly more maintenance issues.
I was comparing it to the equipment used to carry the laser or pesticide. The application of pesticide itself is ~2-3 orders of magnitude larger the laser before you consider manufacturing or transportation.
The laser itself is a rounding error here, what matters is the other equipment involved.
No not poison, but I've been wondering if the lasers could harm anyone's eyesight, human or wildlife. Blind bees? I suspect the incident rate would be extremely low to non-existent.
They're targeted beams. They won't be scanning at eye level. Any pollinators harmed will probably be extremely small in number, and there might even be updates to the computer vision routines that can minimize that harm.
This is going to be very good for the environment once it takes off.
I was imagining in my head kind of real time conversion from sunlight to a laser. Maybe not the practical solution, but I do wonder how doable would that be.
Back of the napkin with GPT assistance seems around 2 square meters needed to power a 150W laser.