You have an overly rosy view of the current situation, and so you are dramatically overestimating how far photovoltaic has to go to improve it, perhaps in part because you have forgotten the existence of 85% of the world's population.
I'm on the grid, and I had a three-week power outage two years ago. Getting to 99.999% would require no power outages for the next 5700 years. Even getting to 99% would require no power outages for the next 3 years, which isn't going to happen; I had an overnight power outage a few months ago. That's because I live in Argentina. Not out in the boondocks — in the capital, with flatscreen TVs, world-class medical care, and subways crammed full of people glued to their Samsung Galaxy Whatevers. A friend of mine in Pakistan isn't even getting one nine; he has a rolling blackout of an hour every few hours. He's at maybe 70% or 80% grid uptime.
I don't understand this myopic focus on "Seattle, London, and really most of Europe". What percentage of the population lives in the US and Europe? 14%. 86% of the population lives elsewhere. How about Tokyo, Seoul, Shanghai, Guangzhou, Delhi, Mexico City, Beijing, Lagos, São Paulo, and Mumbai? Those are the world's ten biggest cities, and by themselves they total nearly the population of the US. As it happens, none of them are in the US or Europe.
Also, as it happens, most of them are at much more equatorial latitudes than most of Europe and even much of the US.
Speaking of rolling blackouts, I lived in California in 2000. At the time, political corruption in the energy markets led to rolling blackouts there, with a total of several hundred hours of blackout at each location (except hospitals, national security installations, and things like that). So California isn't going to reach 99.999% on-grid reliability within the next couple of millennia either, even without any further outages. Reliable electricity is very important to some industries, like semiconductor fabrication. Maxim has a fab there in San José, or at least they did last year. Maybe they'd would be better off depending on solar panels on the fab roof instead of the fickle governance of the California transmission grid.
The upshot of all of this is that transmission-line parity for most of us is going to occur pretty soon actually. It doesn't need to hit 99.999% or 99.99% or 99.9% or even 99% reliability to be an improvement from where we are now. In fact, in Pakistan and in much of the world, even if it's 50% reliable, it's a big help, as long as the times when it's out aren't the same times the regular power grid is out.
Molten-salt solar is certainly capable of some load following, although not as fast as a gas turbine, but it can also handle baseload, because it can run at whatever time of day or night you need. Can it be extended? Well, the materials you need are super common and well understood (we were recently treated to an explanation the other day of how Lavoisier made significant improvements to the handling and production of the salts in question in the 18th century), and the heat engine is typically a closed-cycle Parsons steam turbine, same as in a coal or combined-cycle plant. There's no reason I know of to expect it to be any harder to scale than other types of steam turbines. The only issue is that it will have a hard time producing energy as cheaply as photovoltaic unless we can bring the cost of the machinery way down and its lifetime up.
You mention deserts, I suppose because enormous land area is cheap and the skies are not cloudy all day. But how much area do you really need to generate the electricity we use? Argentine electricity use, for example, is only about 3000 kWh/year/人, or 340 W/人. With a typical panel efficiency of 15% (not 22% like some advanced unijunction panels), a typical capacity factor of 18%, a solar constant of, say, 1000 W/m², and a latitude of 34°36’ (thus a cosine of 82%), this works out to 15m²/人, or 65 000 人/km². The actual population density here in Buenos Aires — the densest and most populous city in the entire country — is 14 000 人/km². Rooftop photovoltaic is plenty to replace the existing rickety power distribution infrastructure.
Pumped water storage might be useful, wind is useful, deserts are certainly going to be immense resources, and cheaper batteries could eliminate the intermittency problem (if they work), but none of those are actually necessary.