Whoops, my bad - I was thinking insolation where I live in the Southwest: clear skies and lowish latitude translate to roughly 5 hours of peak solar equivalency per day, and the solar thermal systems which interest me as a tinkerer can easily approach 100% efficiency (1kW/m^2 at peak). At 16% efficiency that's still about 0.8kWh/day, but not nearly as impressive as 5kWh. On the other hand as you get into more overcast areas further from the equator the appeal of solar thermal increases, and high capacity thermal batteries (aka insulated water tanks) are cheap. Even in northern Montana a DIY thermal installation can pay for itself in a few years, and unless it's replacing wood or geothermal heating that's a big win.
But even PV isn't as bleak as you make out: let's use your number: 0.435kWh/m2/day w/current solar panels. Double that for the high efficiency panels to 0.87kWh. Then halve per-capita energy consumption to get in line with European efficiency:123.5kWh/day. That's still only 142 square meters per person. Three times my flawed estimate, but still not terrible. That's ~44,600km^2, or an area about 30% larger than Maryland, to supply the entire US with all its energy needs. Even with current energy consumption and cheap silicon PV we'd only need an area the size of Missouri to do the job.
And remember: the vast majority of that energy is consumed by businesses rather than individuals, and they are already beginning to roll out solar in a big way, because the $/Watt has already fallen to the point that it's notably cheaper than today's grid electricity over a 20-year amortization period, and businesses are accustomed to dealing with everything in terms of amortized costs. Let the price of fossil fuels keep climbing and the price of PV continue to fall, and it won't be long before PV is cheaper than burning coal on-site in many areas.