First of all, most reverse osmosis system are designed to have a recovery of 20%, that is, for every 1 cup you drink, you throw away 4 cup. That sounds a lot, but you actually don't drink that much water. Baths and lawns are a lot more wasteful. More so is flushing the toilet. Somewhere in the world use sea water or reclaimed water to flush. Also, bottle water, such as the famous Fuji water now widely available in Trader Joe, seems a lot more wasteful, pumping up deep ancient water and bring it over from so far away, in plastic bottles. Also many beverages you drink use RO water anyway. In good restaurants, if you don't want bottled water, they will give you tap water, filtered to taste good. It's worthwhile to include RO to avoid liabilities as bacteria and virus can hardly pass through. It's easy to achieve zero waste. You can feed the water to your irrigation system, or pump it back to the hot water supply.
RO systems used to be very expensive, $1000?, and still are, $500?. Some years back some cheap RO systems appear on the market aimed at the masses at $100. The performance in terms of rejection rate of total dissolved solids (TDS) is 85%, compared to top of the line systems at 98%. 85 is a lot less than 98, but filtering out 85% of what need not be there sounds enough for a try. The DS are basically the hardness of the water, naturally occurring materials. But here we use Colorado river which is really hard. And if something goes wrong for a short while in the public system, softening the impact by 85% isn't that bad. And if you can filter out 85% of the bacteria, that's good enough without costing an arm and a leg. It's not that I couldn't afford expensive systems. But like computers, I don't know what components they put in. You cannot test it, for example, how much lead it rejects. You have to trust them. But I rather not.
But the cheap pumpless systems are a joke. The stated performance is for full water pressure, typically 50 psi (sorry rest of the world). But once the system starts working, the pressure across the RO membrane falls. These systems typically stops when the tank pressure reaches 30 psi, by this time the pressure across the membrane is only 20 psi, a fraction of the stated performance.
Typically small manufacturers don't know what exactly happens as the operating points drift away from the optimum point where all the components are guaranteed. But as long as the final TDS reading looks good, they are selling it. Typically, the rejection rate maintains at low pressure, at the expense of wasting much more water! Not just 20%, but 10% and less.
To control the pressure, you need a pump. But pump's able to deal with 50 psi are not toys. Just a booster pump cost more than a cheap RO system. The alternative is to use an open tank and let the RO water flows into the tank slowly. So you are fortunate if you are a fish. For aquarium applications, the water is as good as any top of the line systems. But without a pressurized tank, you either go fetch water from under sink, or use a delivery pump. So a pump again, though at lower pressure the pump is cheaper, but not by a lot needing NSF testing. And you start to worry about the air quality, and air borne bacteria. Until the permeate pump, which is a lot cheaper without electricity and at lower pressure. Basically it's a weak hydraulic pump, pumping filtered water into the pressurized tank. With a pump, the pressure of the tank is isolated from the membrane. So the membrane is seeing the full water pressure all the time.
I'm going to describe safe materials and the correct design of an RO system. Indeed if there are not so many one man company selling put together systems, I may start selling too.