You over estimate the decay heat in a kg of spent fuel. According to here [1], a 1GW (1000MW) plant uses 100 Metric tons of fuel, so 1 kg is responsible for 1G/100K or 10Kwatts and the decay heat of spent fuel is .1% of its nominal power level so 10K * .001 or 10 watts. If you dump 10 watts into your water heater it will get hotter and depending on the heat insulation of the tank it might even get to boiling. Although typically I would expect a tank to lose close to 10W into the ambient air once the water was at temperature.
No argument, though it's hard to envision a scenario where [1] is not a predetermined eventuality for any setup that would be laughably inadequate or economically impractical.
It doesn't follow, fuel (whether spent or not), is density and dissipation question. Each kg of fuel is generating heat, so more fuel in a smaller space means more heat in a smaller space. Fukishima's spent fuel ponds had both spent and fresh fuel in them (reactor 4 was undergoing maintenance) and there was a lot of it. Spread that fuel out across say a hectare pool of water that is 5 meters deep and it stays out of trouble.
One disposal plan called for mixing the fuel with silica to create glass bricks. By spreading the fuel mixture to a density that was low enough to allow the ambient air to cool it, the bricks are shelf stable for thousands of years and not a radiation hazard (less of a risk than even natural uranium deposits you might come across).
Many disposal plans were like this, encapsulate the spent fuel into a container that withstands (and blocks) the radiation while conducting enough heat to avoid changing the properties of the material over the lifetime of its storage. Glass and austenitic stainless are both good candidates for that material.
The fact that spent nuclear fuel must be kept in water at all until it is sufficiently decayed to be put into dry storage still safely away from humans means it is by definition not fail safe. Water evaporates and if you lose the ability to dissipate heat or absorb ionizing radiation, "you're in deep shit" would be a colossal understatement. How comfortable would you be that something on your property or in your home is one water outage or tank failure away from irradiating your family and probably a bunch of your neighbors. I imagine the hazmat team won't be happy to be invited to your housewarming either.
If it requires cooling or radiation-blocking-immersion to be safe, it absolutely has to be monitored by trained personnel 24x7 who can also immediately and reliably take necessary measures when something goes wrong. Anything less than this is a non-starter. Nuclear reactors work because they centralize risk to where it can be managed by a team of knowledgeable people; A 1GW reactor does not have the same risk profile as 1000 dispersed 1MW reactors.
But that was my point, spent fuel doesn't have to be kept in water at all, if it is sufficiently dispersed. Take a spent fuel assembly remove each of the fuel rods and lay them down on the ground and you're done caveat two things, one they emit alpha, beta, and some gamma radiation and two they warm their surroundings. So there always exists a solution to the equation of heat generated over time = heat absorbed by ambient air keep the total temperature rise below that of the safe temperature of the cladding.
If you lay a spent rod out on the ground in the air, it may glow (would depend on the amount of material in a single rod) and you wouldn't want to approach it without protection (again more material more radiation) but it would be fine.
The reason people put these things in water is two fold, one is that water is a great transporter of heat and you can generate almost an arbitrary amount of heat in a small space if you have enough water flow to carry it away (ask any highend gamer PC aficionado :-)) and water absorbs neutrons, which if they were not caught and hit another fissile atom nearby slowly enough might spit it, or a non-fissile atom might become an unstable nucleotide (which is going to eventually emit an alpha or beta particle)
If the material is clad in a neutron opaque material and can transfer the latent heat to the surrounding atmosphere or fluid at a rate that keeps the system temperature under the breakdown temperature of the containment, then it will be stable "forever".
Yes, it takes a year or two for the short lived nucleotides in a fuel rod pulled from a reactor to decay, but once it has, you can safely store this stuff.
In terms of why we haven't, most of the issues around storing high level nuclear waste arise because third parties do not want to encapsulate permanently this material, because doing so would make it uneconomical to recover. And in that calculus they know its "easier" to reprocess spent fuel for bomb grade uranium and plutonium than it is to build a breeder reactor to make it.
You could, but it would be less than 10W of electricity, so not enough to power a lightbulb. The water tank is insulated, so the water just heats up slowly until it reaches boiling point. That doesn't mean it's a powerful heat source.
[1] http://www.nucleartourist.com/basics/hlwaste.htm