I would really like to see some numbers coming from the little box with wires. There are lots of issues with accelo-generators that have quashed this idea for, oh, the past 150 years.
The basic idea works -- and it works well on very small scales (self-winding watches) and very large scales (tidal generators). The keys there are minimal output energy requirements (wristwatch) or extremely powerful energy inputs (a few million tons of moving water).
Work is work: a force applied over a distance. To reap watt-hours from an accelo-generator, you need a very heavy weight, or a very large input acceleration, or a really long stroke. Preferably all three. One horsepower (~750W) is equal to lifting 550lbs one vertical foot in one second. Let's say the weight in the box is 15kg (33#), the pendulum stroke is 8", the acceleration under frisky conditions is 0.5g, and the conversion efficiency factor is 50% (that's extremely generous). Back of the envelope calcs suggest 2.75 watts out of such a device., or 0.2A.
Now where do you put it? Boat motion is greatest outboard and at the ends -- but you don't want heavy stuff or accelerated masses on the ends of your boat, or outboard. There's the issue of hysteresis: the device is going to lag behind the phase motion of the boat, esp. in erratic seas, cutting efficiency further. Most boats spend most of their time in a marina or harbor, chosen and designed specifically to minimize motion from waves.
It's a neat idea, I guess; but I'm always deeply suspicious of fractional horsepower electrical generating devices, especially on boats. How about doing what inventors used to do: Build the magic box, put it on a sailboat, show us third-party verified power outputs when attached to an actual load, and then go looking for someone to capitalize & manufacture it? I'll be thrilled if it works, might even buy one, but til then it's just mo' RE vaporware.