I expected some stupid metaphor, but he title is technically true, but let's do the math:
The concentration of Hydrogen in air is 0.0000038% by mass. To get 2g of Hydrogen you need 50000000g of air. That is approximately 40000000 liters of air, that are 40000 m^3 of air.
With 2g of Hydrogen, you get 286kJ = 286kWs. Let's assume a 100% efficiency conversion that is very optimistic, and let's round it to 300kWs. That's enough to turn on 20000 lamps of 15W for 1 second.
So to keep one light on you need to pass 40000/20000 = 2m^3 of air per second. If it's a light, I expect a system/fan of a similar size, like a PC fan. Googling for a while, I found that a typical PC fan has only .2m^3/s.
So you need more than 10 PC fans blowing air in the system to turn on a lamp, assuming they don't need electricity.
Another possibility is to use the natural wind. Cherry-picking some numbers from https://en.wikipedia.org/wiki/Wind_turbine#Records a turbine with a swept area of 38000 m^2 and a wind speed of 10m/s gets 400000m^3/s of air, that is optimistically 20g of Hydrogen per second, and optimistically 3000kW = 3MW. That's less than the 12MW of the turbine, and assuming a 100% of capture of the Hydrogen and also a 100% of conversion to electricity is too optimistic.
Based on the article, it sounds like initial applications are possibly worn devices or other small devices that use rechargeable batteries. The mechanics of being worn, or possibly even creating air convection by lining a heat sink with this stuff, might allow it to access sufficient airflow.
That said, it'd be interesting to find larger structures that could take advantage of natural convection where wind turbines or solar panels might not be terribly suitable, but that's probably a bit of a pipe dream; I doubt this stuff can both work and be exposed to challenging environmental conditions (dust, snow, freeze / thaw cycles, rain, animals scurrying around on it, etc).
A watch battery has 1.55V and 40mAh. That's 1.5 * 40 / 1000 * 3600 J = 216 J. How long does they last? 1 year? That's 216/365/24/3600 W = 0.0000068 W ~= 0.000007 W, i.e. 1/2000000 of my lamp.
The lamp needs 40000000 liters of air per second, so a watch needs 2 liters of air per second. (That's half a gallon.) One person breath like 1 liter each time, 15 times per minute, so blowing your wrist you get only .5 liters per second. To get 2 litters per second you must be running or moving your arms like crazy. And still I guess you will not have enough air flow.
The concentration of Hydrogen in air is 0.0000038% by mass. To get 2g of Hydrogen you need 50000000g of air. That is approximately 40000000 liters of air, that are 40000 m^3 of air.
With 2g of Hydrogen, you get 286kJ = 286kWs. Let's assume a 100% efficiency conversion that is very optimistic, and let's round it to 300kWs. That's enough to turn on 20000 lamps of 15W for 1 second.
So to keep one light on you need to pass 40000/20000 = 2m^3 of air per second. If it's a light, I expect a system/fan of a similar size, like a PC fan. Googling for a while, I found that a typical PC fan has only .2m^3/s.
So you need more than 10 PC fans blowing air in the system to turn on a lamp, assuming they don't need electricity.
Another possibility is to use the natural wind. Cherry-picking some numbers from https://en.wikipedia.org/wiki/Wind_turbine#Records a turbine with a swept area of 38000 m^2 and a wind speed of 10m/s gets 400000m^3/s of air, that is optimistically 20g of Hydrogen per second, and optimistically 3000kW = 3MW. That's less than the 12MW of the turbine, and assuming a 100% of capture of the Hydrogen and also a 100% of conversion to electricity is too optimistic.