Shellenberger, as president of a pro-nuclear interest group Environmental Progress, has dedicated himself to keeping current nuclear plants running around the world and advocating for building new ones. Reading over the group’s materials, I am reminded of the grand technological visions that got us into the nuclear game in the first place. Nuclear energy was originally promised to provide energy “too cheap to meter” and an atomically powered utopia. Ecomodernist environmentalists like Shellenberger see similar nuclear promise, namely as a Deux Ex Machina for climate change, if his cited estimates of its carbon footprint are to be believed.
It is important to note that the economic (and safety) estimates that drove the first nuclear bubble proved to be spurious. Largely based on theoretical projections made in the absence of real experience, they proved too optimistic and utilities were stuck with underperforming reactors that were completed years, if not decades, behind schedule. Nuclear energy has only become moderately financially competitive as decades-old investments have been paid off. Given these past mistakes, we should be worried about becoming too enchanted with a similar, albeit ecomodernist, nuclear day dream.
Shellenberger and others might prove correct regarding nuclear’s carbon footprint, or they might not. Getting an accurate life-cycle assessment is challenging for most technologies, and the carbon picture for nuclear is made worse when including the processing of uranium ore and decommissioning of reactors. Considerable uncertainties emerge when recognizing that we do not really know the carbon costs of building and maintaining long-term storage sites. Will nuclear look so advantageous if we massively expanded our use of it and when we finally gain some practical experience with stewarding the waste? Even worse, we might not actually learn the full environmental costs until it is too late, inadvertently committing ourselves to a net environmental negative for ten thousand years. And that’s not even accounting for the carbon footprint of potential nuclear conflicts made more likely, because expanding nuclear power across the globe would almost invariably lead to more nuclear weapons programs.
The ecomodernist position acknowledges little of these complexities. Shellenberger seems to know just enough about nuclear energy to paint himself into a technologically conservative corner. In an article for Forbes, he cites the financial and scheduling woes of new designs like the AP1000, which promised more passive safety features to overcome some of the light-water reactor’s inherent design problems, as evidence that nuclear innovation writ large is costly and should be strictly limited. His article dwells on Hyman Rickover’s experience with developing the light water reactor and statements about the challenges that arise when translating experimental designs to the real world. Shellenberger advocates that we mostly stick with a slightly modified, standardized version of decades-old designs—to be produced in massive numbers—ostensibly because it fits with the ecomodernist demand that we rapidly expand nuclear energy to combat climate change. Moving to a different design would slow down that process.
The costs of innovation, however, is only sometimes and partly due to the radicalness of design changes. Also pertinent are the capital intensity of the technology, speed of feedback about errors, dependence on specialized infrastructure, and the process for scaling up. It is about the whole sociotechnical system of innovation, not just the technology itself. The challenges that nuclear innovation faces today is partly due to the shape of the system put in place decades ago.
The costliness of pursuing potentially safer reactor designs—or one’s that may have a better outlook in terms of waste or carbon emissions—is partly caused by previous generations of nuclear advocates plunging ahead. The currently dominant light water reactor became the de facto standard largely before we learned much about its relative benefits and a drawbacks for producing commercial energy. We still do not know much about the alternatives. Options became foreclosed because true believers chasing nuclear dreams, pursuing market dominance, and/or wanting to beat the Russians acted single-mindedly, constructing dozens of light-water plants and scaling them up by a factor of six before people started to recognize their flaws and the bubble burst.
As a result of this nuclear energy bubble, alternative designs face unfair comparisons with a light water standard that received billions in early investment and subsidy, among other advantages. Alternatives designs are further stymied because established infrastructures, educational programs, and regulatory regimes are not designed for them. And people are more risk-averse with regard to developing alternatives, lulled into complacence by the knowledge that a working design already exists. Put together, this leads the light water reactor to be the QWERTY keyboard of nuclear energy: a suboptimal design that is nevertheless too entrenched to be altered or dispensed with.
People who have studied how promising innovations turn into expensive technological failures warn about the risks of locked-in or otherwise inflexible technologies. The process of standardization and narrowing of alternatives that Shellenberg sees as so beneficial financially is only desirable once one really knows the ins-and-outs of competing designs. Ensuring such diversity is no doubt expensive, but expenses can be reduced by ensuring that development is appropriately paced and scaled up gradually. Consider how NASA-led wind turbine development in the United States was a costly flop: Engineers scaled up to megawatt sized turbines that failed early and catastrophically, providing little to no guidance for commercial builders. The Danish, on the other hand, supported a decentralized process by a diverse set of builders. The size of the turbines grew gradually over the course of a decade, resulting in Danish turbines being the most reliable in world without millions in wasted on inefficient investments. Note that this is the complete opposite of the process that Shellenberger advocates: decentralized rather than centrally controlled, small and gradual rather than deployed at grand scales, and diverse and open rather than standardized from the get go.
Realizing such a decentralized and gradual process with nuclear energy is challenged by a number of factors—no doubt securing nuclear material being one of them. Perhaps the biggest barrier is size expectations. Any moderately novel reactor design is going to run into major unanticipated problems when it is expected to produce electricity at the scale of hundreds of megawatts. This well-established pattern should give us pause. Can we create a context where intelligent incremental nuclear innovation can occur? If we cannot, then perhaps for all intents and purposes nuclear energy is a pathological technology. If its very character prevents a learning-focused innovation process—especially with respect to safety and environmental impact—maybe it is not worth the cost and effort, especially given that energy reduction is so much more cost effective and freer of undesirable consequences.
The coach of my university’s rugby team instructs players that “Slow is smooth, and smooth is fast.” Behind what would seem to be an apparent contradiction in terms is a recognition that panicked decisions are often more costly in terms of time (and everything that comes with it) than being self-conscious and deliberate. Another popular rugby saying is “Don’t shovel shit” This means if you receive a crappy pass, don’t try to pass it on to the next guy. It will only make things worse.
The seriousness of the problems of global climate change should not be viewed as an invitation to make hasty decisions about nuclear; such choices could make the predicament of future generations even worse. Without an emphasis on deliberate learning, the risk of doing so is very high. Moreover, there is no requirement that we remain committed to our nuclear inheritance. The sunk cost of experience, dedicated infrastructure, and other choices that now make current light-water designs seem like a good deal was originally the sociotechnical equivalent of a shit pass. We do not have to shovel it on to the next generation. We can take the opportunity—as more and more of our plants reach the end of their lifespan—to reassess the situation and adjust our line of attack, proceeding in full recognition of our limited knowledge.