Sun and wind vary during the day and the year (obviously there is no sunlight at night, and less of it in winter than summer), completely independently of needs. Nowadays electricity production is tuned to balance consumption, but sun and wind cannot be tweaked in this fashion. It will be the consumption side that, in the future, will have to match the production, rather than the other way around. Smart grids and peak demand management will play key strategic roles in matching electricity production and consumption.
Going from 4% to 5% of renewable energy (photovoltaic and wind) is easy. Going from 4% to 40% is a completely different story. The cost of increasing renewable electricity a little is the cost of making and setting up solar panels and wind turbines; this is the often-quoted price of electricity from this or that source. But going from a token bit of renewable electricity to a sizeable chunk, say multiplying by ten, will not cost ten times as much. Economies of scale and technological progress will reduce the price. But, on the other hand, a need for a suitable infrastructure will emerge, adding to the cost. (It will also drain political will and public motivation; in countries where these are rare but nuclear power is plentiful, especially France, the switch out of nuclear electricity would be too much of a drain.)
Currently, small-scale producers of photovoltaic (PV) electricity piggyback on the grid: they sell and buy electricity when it is convenient for them (so that even those who, in total, consume as much as they produce are not autonomous), and they produce as much as possible, even if it means maximizing production when no one wants it. This is possible only because it happens on a small scale.
It is possible today to buy and use an all-electric car. Yet, for most people this means charging it at home, for there is no other place where to charge it. This (on top of issues of price, range, etc.) is a clear limitation for potential users. Of course, a network of charging stations can (and certainly will) be built. What this means is, as with large-scale renewable electricity production, that an infrastructure will be needed for all-electric cars to be mainstream. Financially, the cost of switching a million cars from petroleum to electricity is not just a million times the price difference; it is to an extent the cost of the necessary infrastructure. And in practical terms, it means that making the switch is not a purely individual decision.
Electric cars are often touted as synergetic with the electric grid: PV and wind electricty produced at the wrong time could be used to charge car batteries (along with other storage), upon which one could draw when consumption is high and production low. But for this system to work, we need PV and wind production on a large scale, efficient energy storage on a large scale, a smarter grid, lots of electric cars (and therefore a battery-charging infrastucture), etc. This sure looks good on the face of it; but how do we get there, given that this requires setting up not one but several infrastrcutures?
Recycling and composting too require an infrastructure (albeit at a smaller scale than the previous two), to pick up or receive the recyclable wastes and then to treat them. But this infrastructure is more common nowadays than for the smart electric grid or electric cars, so we tend to forget about it. But when travelling to a place where multiple trash cans are not the norm, I feel powerless: I do want to recycle my glass jar or my newspaper, but there is nothing I can do. Likewise, I normally compost organic waste, so when I travel I never know what to do with my apple cores. All the good will in the world is for naught if the proper infrastructure is not in place.
These three cases are in contrast to architecture and agriculture, where large-scale change requires no large-scale infrastructure.
The energy efficiency of buildings has been improving over the past forty years (in earnest since the 1973–74 oil crisis). But this happened a building at a time, without the need of a far-reaching infrastructure revolution. In the case of the energetic efficiency of buildings, the bottleneck was not the need for a new infrastructure, but rather the lack of motivation. Amory Lovins' house has enough insulation to "eliminate the building's heating system and thus avoid the entire capital cost of furnace, ducts, [etc.]". And this was thirty years ago (and in a cold place). A house could be both efficient and economical, and there is no need to wait for infrastructure.
Growing food without synthetic fertilizers and without chemical pesticides is a way of reducing the environmental impact of agriculture. It can be done at any scale — there is no need to wait for a critical mass that would justify some agricultural infrastructure. This is not to say that governments cannot do anything: they could internalize externalities (such as global warming, pollution of water by nitrate run-offs, droughts downstream from excessive irrigation, etc.) and of course by reducing their subsidies to harmful practices. But these are not required, the main reason why the majority of food is not grown more sustainably is economic: if organic food were cheaper than conventional production it would soon sell much more than the latter, whereas a price drop (or heavy subsidies) would not suffice for a large-scale conversion to electric cars in the absence of a charging infrastructure.