People are always asking Bill Gates if he's really read all of Vaclav Smil.

Smil is a professor at the University of Manitoba. His bio says he does “interdisciplinary research in the fields of energy, environmental and population change, food production and nutrition, technical innovation, risk assessment, and public policy.” He’s a wide-ranging thinker. He's published 36 books. One is about the Japanese diet. Gates doesn't recommend it.

Smil has been particularly influential in his work attempting to quantify the energy needs of global civilization, and the likely energy yields of various renewable alternatives to fossil fuels. It’s a dizzyingly comprehensive project. Smil is attempting to put numbers to the human use of energy across all of history (and some of pre-history) to enable meaningful comparisons between various phases of human civilization.

Smil focuses on “energy transitions,” epochal shifts in how people have harnessed and used energy. He thinks with some good high-level analysis, understanding those transitions will help us make more realistic plans for the necessary shift to renewables.

The stakes are clear. The planet is hurtling towards a 2 degrees Celsius rise in average temperatures, and likely beyond. To mitigate climate change, we need to significantly curb global carbon emissions. At the same time, standards of living are tightly correlated to energy consumption (to a point); there are billions of people who'd like to lift themselves out of poverty, and the best known methods available for supplying them with power come from fossil fuels.

Smil is convinced that a transition to renewable energy is inevitable in the long run. But in the short term, he thinks it will take far longer and be much more disruptive than most green-energy proponents would like to admit.

On the optimistic side, you have advocates like Al Gore. In a 2008 speech, Gore said that “enough solar energy falls on the surface of the earth every 40 minutes to meet 100 percent of the entire world's energy needs for a full year,” and called for the U.S. to generate all of its electricity from renewables by 2018. Smil argues that these kinds of pronouncements ultimately do us a great disservice by drastically overstating the clean energy that's actually available to harness productively, while understating the work, time, and investment that will be needed to transform our energy infrastructure. Into this debate, steps Gates, Smil's number-one fan, with a $2 billion pledge to invest in research and development to accelerate the transition to renewables.

Reading Smil is a wild ride. He dances up and down scales, moving from broad global trends to very specific local examples to illustrate his arguments. In Energy: A Beginner's Guide, a discussion of the energy costs of a human picking an orange up off the floor and putting it on a counter is separated by mere pages from an estimate of the energy expended by an erupting volcano. Soon after, you are learning about the (hilariously inefficient) photosynthesis of plants and comparing the productive capacity of the biosphere with likely constraints on our ability to harness solar power at planetary scale.

To keep up with Smil, let me introduce a few of his basic concepts: To begin, we need to understand what he means when he talks about resources and “prime movers.”

First, the resources. Pretty much all available energy on the earth comes from energy radiated by the sun. Straight sunlight offers heat and light. Some of that is captured by plants, which can in turn be eaten or burned for light and heat. The plant eaters (human or otherwise) can be put to work or eaten. Some of the sunlight evaporates water, which rises into the atmosphere and then falls as rain. Some of this rain lands on very high places and converges as gravity pulls it into streams and rivers that run towards lakes and oceans. The mechanical energy offered by running water is essentially stored sunlight. Meanwhile, differences in temperature because of the uneven distribution of sunlight (owing in part to the shape, tilt, and rotation of the earth) causes air to circulate, meaning winds are stored sunlight too. As are waves (which are driven by winds). And, of course, fossil fuels are reserves of stored sunlight, derived from biological depositions which have been accumulating for millions upon millions of years.

The exceptions to the sunlight rule are: geothermal energy, which comes from the very hot core of the earth (often in the form of volcanoes); tidal energy, which is the result of water interacting with the gravity of the earth, moon, and sun; and nuclear energy, which comes from either forcing together or breaking apart atoms. Some of these resources are renewable, but at the moment, the dominant suppliers of energy to human civilization (the fossil fuels) are not.

Second, there are Smil’s “prime movers,” which he defines as “energy converters able to produce kinetic (mechanical) energy into forms suitable for human use.” For most of the time that there have been humans on earth, the best prime movers have been people. We eat food, we convert the chemical energies stored in plant and animal flesh into muscle power and we use that to do mechanical work.

Some numbers will give a sense of how prime movers have increased in power over the course of history. Remember that orange lifted to a counter? If you expend that effort over a second, that's 1W (a watt) of work. Smil calculates that the average healthy human can sustain 60W–100W of work throughout a working day. At some point in prehistory, people started yoking domesticated animals (250W–800W depending on the breed). Then came sails, then a few thousand years later, waterwheels (2,000W–4,000W in medieval times) and then a thousand years after that, windmills (1,000W–10,0000W in 1900).

Things really kicked off with the invention of the steam engine in the 1700s—the first prime mover powered by fuels (100,000W in 1800; 3,000,000W in 1900). This was followed by the steam turbine (75,000W in 1890; 25,000,000W in 1914). The prime-mover revolution is rounded out by the internal-combustion engine in the later half of the 1800s and the gas turbine in the 1930s.

Smil is concerned with the series of transitions that have occurred throughout human history, both in terms of resources and prime movers. These transitions are somewhat interrelated, but not completely. For example, you can run a steam turbine off of wood, coal, or nuclear power so a transition between those resources does not necessitate a change in prime movers. On the other hand, you can’t feed an internal-combustion engine with wind or wood. At the moment, all of our best prime movers rely heavily on fossil fuels.

Historically, Smil says, we can trace the transitions of prime movers to generational developments of technology. His favourite example is liquified natural gas. The technique to liquify gas was first discovered in 1852. The road from that invention to a global-scale infrastructure of extraction, refinement, transportation, and use in a true marketplace took 150 years. Smil recounts similar timelines from invention to commercialization for other fuels and prime movers.

Those transitions have also been heavily dependent on the energy infrastructure that came before. The age of steam was not possible without human and animal work to mine the coal and build the machines. Even now, the wind turbines we look to to help us escape fossil fuels are steel towers (you make steel in coal-fired blast furnaces) topped by plastic blades (which comes from petroleum), installed by (gasoline-powered) construction equipment. A wind turbine is a “pure expression of fossil fuels,” said Smil during a 2013 lecture at the Perimeter Institute.

So, while Smil agrees with pretty much everyone else that the next big energy transition is from nonrenewable to renewable resources, he is cautious about the timing. At one level, the change is plainly inevitable. There will come a time when non-renewable resources run out, and Smil says it will be advantageous to transition off of fossil fuels long before then, to avoid climate change.

In this, Smil is no different from countless energy advocates from Greenpeace to Al Gore to T. Boone Pickens. Where he does differ is in his opinion about how quickly it can happen. Where Gore calls for a complete conversion to renewables in 10 years, Smil thinks the transition will take generations.

Smil's books devoted to this argument are awash in numbers, statistics, estimates, and calculations. I can’t walk you through them all, so checking his math is left as an exercise to the reader. But some sketches of the high-level objections give an impression. In particular, the final chapter of Smil’s Energy Transitions: History, Requirements, Prospects reads essentially as a design brief for the problem that Gates is trying to tackle.

The barriers to total conversion—much like the problems that plague our energy infrastructure—are a funny mixture of policy, technology, infrastructure, and physics. For example, the possibility that nuclear power might take up any of the load in the U.S. seems extremely low, given that no new plants have been built since the 1970s. That’s not a physics problem, that’s a policy problem. Infrastructurally speaking, nuclear is not that different from coal or liquified natural gas. Fuels are brought to a site, where they are consumed while giving off enormous amounts of heat that runs a steam turbine which generates electricity. We know how to make nuclear reactors and hook them up to the grid, but Americans just don’t want to do it anymore. And there are good reasons for this! But compare that decision to the one made by France, which went all-in on atomic power and generates about 75 percent of its energy from nuclear reactors.

As far as converting to wind and solar, Smil sees much bigger technological and infrastructural hurdles. A switch to renewables means a transition in terms of both resources and prime movers. The character of renewable resources is fundamentally different from that of fossil fuels. Where fuels are highly dense stores of energy and relatively easy to reliably transport, the renewables are characterized by the highly fickle ebbs and flows of nature. Some days are sunny, others have clouds. Some days are windy, others are quiet, still others are too windy to safely run the mill.

The exception is biofuels, but remember how hilariously inefficient plants are at converting sunlight to biomass? Converting that biomass into fuels adds even more inefficiencies and you end up needing to convert huge swaths of the planet's surface over to fuel growth. Smil thinks it's completely unfeasible.

Which brings us to Smil's last key concept: energy density. Energy density is sometimes used to discuss the capacity of volumes of batteries and fuel. Smil is interested instead in measuring energy per unit of the earth’s surface. He uses the figure as a means to try to compare the various means of producing energy and the demands for using it.

So, to measure the energy density of coal, you look at how much energy you get from burning coal and divide that by how much of the earth’s surface needs to be given over to coal production to get it. Smil’s calculations try to take into account all kinds of complications involved in energy infrastructure. There’s the extraction point (in Smil's method, an open-pit mine has worse energy density than a hole drilled in the ground), the transportation networks, the space needed by any refineries, the area of land disturbed around these areas, the possibility of damage from spills, etc.

Every fuel transition (wood to coal to oil to gas) has involved moving to higher densities. We have benefited every time from more energy per area of land given over to energy infrastructure. Renewables buck that trend. Once you account for the inefficiency of conversion and the irregularity of generating at full capacity, the energy density of wind or solar installations is far below that of oil and gas. They simply take up a lot of space. Because the best way of mitigating the irregularity in how they generate power is to create interconnected grids, an energy regime based on wind and solar needs to lay a lot of power lines through a lot of jurisdictions and permitting regimes. Physics meets infrastructure, and policy. Renewables are simply more diffuse.

At the same time, energy consumers are becoming more concentrated. Economies of scale bring higher-intensity manufacturing to smaller areas, and people are becoming increasingly concentrated in denser and denser cities.

“Mass adoption of renewable energies would thus necessitate a fundamental reshaping of modern energy infrastructures,” Smil writes. We'd go from harvesting energy from concentrated sources and diffusing it outwards, to gathering energy from diffuse sources and concentrating it inwards towards relatively few centers. This is, fundamentally, a very different way of organizing how we use land.

This is not impossible, and in the long run it is probably inevitable. But we underestimate the effort required and changes that will be necessary at our peril. The switch from wood to coal ushered in industrialization which completely upended social-structures and human relationships all over the world. The rise of oil transformed geo-politics, turning some countries into energy superpowers overnight. No one knows how deeply our society might be transformed by the transition to renewables. Or whether we'll be able to do it fast enough.