Ecological Prospects for Computing

This piece was originally written in the first quarter of 2018, and published in issue #10 of The Recompiler, in print as of November 2018 and online as of 2019.

Introduction

Reports now estimate that Bitcoin uses over 65 terawatt hours of electricity per year, over 0.3% of total global electricity, with the first peer-reviewed study projecting its usage to reach 0.5% by the end of 2018. My first thought is: how could you use so much energy to do so little?

Bitcoin is at its core a system for decentralized and verifiable updating of transactions and balances on a globally shared ledger, with various enhancements to the core protocol such as smart contracts. I do not deny that pseudonymous electronic cash without the need for central banks benefits some people, but as an anti-racist and anti-imperialist socialist, I will not accept that the satisfaction of these desires deserves more electricity than Nigeria (population 186 million).

This is just one of a countless number of examples of the grossly unjust allocation of all types of resources under capitalism. Consider that there are six times as many empty homes in the United States as there are homeless people, or that at least 20% of food is wasted worldwide while close to a billion people go hungry, according to United Nations statistics.

Like issues of housing and food, the question of the distribution of energy and computing resources is embedded in a larger societal context, a principal feature of which is the catastrophe of climate change. This essay places front and center the larger political struggles over the allocation of resources in our unequal and oppressive setting, surveys the ecological impacts of several actually existing technologies, and imagines possible futures for computing in light of the above.

Foundations

To begin this discussion of ecological prospects for computing, it is necessary to start with the fact that climate change is a racist crisis. This can be observed in the disproportionate impact of carbon pollution and global warming on oppressed peoples compared to its effects on the nations, organizations, and individuals that are actually most responsible. Negative effects are most clear upon denizens of the Global South, but indigenous peoples, Black people, and other people of color and poor people living under the U.S.-led imperialist rule of Europe and its settler colonies are also subjected to environmental colonialism. (These categories are already indistinct and will only become more so as the numbers of migrants and climate refugees continue to climb.) The era has been named the “Anthropocene,” as though humans writ large are responsible for this unfolding disaster. I prefer the alternative term “Capitalocene,” which instead lays responsibility where it belongs.

Besides my deep-rooted distrust of imperialist countries and businesses, I have found it necessary to apply related critiques to the analysis of other technologists on this topic. For example, “What Can a Technologist Do About Climate Change?” has been an invaluable resource, but I find it utterly lacking in political perspective. Regardless of intention, failure to take an explicit stance against the system which has produced this nightmare scenario almost inevitably results in implicit support for the oppressive capitalist status quo, for example in this section about funding for climate change solutions:

Available funding sources, and the interests and attitudes of the funders, have always had an enormous effect on what technology comes to exist in the world. In a time of crisis, it’s the responsibility of those holding the capital to sponsor work on averting the crisis. That is not happening.

The first sentence I agree with completely; you can read some of my views on the pervasive manifestations of imperialism in technology in my prior writings on the colonial biases of a software library I used in an art project. The second sentence, on the other hand, I consider misguided at best, for it lacks class analysis and fails to ask the basic question of who has the motivation to make what types of changes. Malcolm X taught us that “You can’t have capitalism without racism.” Assata Shakur warned in her autobiography that “Nobody in the world, nobody in history, has ever gotten their freedom by appealing to the moral sense of the people who were oppressing them.” These lessons echo through Black history, from the Haitian Revolution and earlier to contemporary abolitionist struggles against the continuation of slavery in the guise of police and prisons.

With this grounding in the thoughts and struggles of my honored ancestors (and elders still with us!), it would be absurd to appeal to the same capitalists who have actively lied about climate change for decades and deliberately schemed to bring about this very situation for the sake of their own profit and power. Yet such appeals, which I call liberalism (among other things), are pervasive in technical circles and extremely harmful. Liberalism leads many well-intentioned people to waste their efforts on reformist strategies that have never worked and never will. Given the interests of the ruling elite that primarily controls society’s resources, there should be no surprise that investment in solutions to these problems falls so far short of what is required: pillage of the Earth and impoverishment of masses of people are the requisite sources of capitalist wealth and power. Building an equitable society in ecological harmony necessarily entails confiscating and redistributing the ill-gotten wealth of the capitalist class.

In the final section of this article, I will share my views on how to achieve such a reallocation of resources, and highlight some of the work before progressive technologists to build and strengthen community-based technology. But first, we must discuss some of the technologies that exist now.

A programmer’s perspective

My study of computer science and previous employment as a software developer trained me to reason about the resource efficiency of programs in terms of “time” or, more accurately, CPU cycles and “space” or, more accurately, memory and storage usage. I honestly cannot recall reflecting on the ecological impact of a program I’ve written, at least not in terms of physical resources like electricity. In the paper that originally introduced Bitcoin, despite discussion of who controls compute power within the network, there is no evidence that the author(s) factored environmental concerns into the design of Bitcoin, either. Such inattention is not universal among software engineers, but it is ubiquitous and highly consequential.

While I still consider programming to be the closest thing to magic—conjuring new concepts and applications into existence merely by typing the proper incantations—the unalterable physical reality of computation is that our abstract fantasies can only be summoned into existence by a sacrifice of scarce physical resources. It should by now be clear to all of us that when designing large-scale distributed systems, ecological considerations are in fact highly relevant.

I am far from being the first to consider questions such as these. Let us examine some existing work at the intersection of computing and ecological sustainability.

State of the art

The computing supply chain

In discussing the connections between ecology and computing, it is fitting to begin with where and how hardware manufacturers obtain the raw materials required to produce computers and what happens before the end user acquires the device. This topic is worthy of volumes, but I can treat it only briefly here.

A typical cell phone contains numerous metals (and metalloids) including copper, iron, nickel, aluminum, lead, gold, and silicon; various plastics are also major components. Similarly composed items include other consumer electronics, the computers found in large scale data centers, and military equipment. Mining industries have been notorious for centuries for extreme environmental destruction and exploitation of workers, and this continues to be true in the heavily African and overwhelmingly Global South nations at the base of high-tech supply chains. Once workers from places like the Democratic Republic of the Congo, China and Brazil have extracted the raw materials from the Earth, typically under appalling conditions with severe negative health consequences, shippers transport the materials to sweatshops located primarily in Asia, where a predominantly female workforce assembles them.

Studies indicate that even suppliers of mobile phones receive only a small fraction of the value derived from their use; thus, the compensation received by the workers who actually create this value in the first place can only be even less. Such plunder of the Global South for the benefit of well-off classes in imperialist countries is absolutely fundamental to capitalism today, as it has been for centuries. Perhaps few engineers consider such a topic when sitting down to work, but it is the indispensable foundation for the platforms we have designed, such as those discussed in the following sections.

Cloud computing

Large corporations such as Google and Amazon provide “cloud” platforms that allow countless organizations around the globe to use existing computing infrastructures rather than create their own. By making use of various “pay-as-you-go” models, organizations can realize cost savings by avoiding large up-front investments in systems that they rarely, if ever, use at full capacity.

In theory, the intense power consumption of the data centers Google, Amazon, and others operate could be offset by ecological benefits from reducing the total amount of hardware required to perform all of society’s computing. But in practice, the connection among conservation in financial, computational and ecological resources is tenuous, to say the least, even before considering questions of who determines the content of society’s computing.

Google and Amazon both actively support U.S. military interventionism. In addition to being a primary enforcer of ecocidal capitalism, the U.S. military is the largest polluter in the world. So I consider deceptive statements by Google and Amazon that present themselves as progressive and committed to environmental sustainability. Even if Google’s and Amazon’s data centers ran completely off of sustainable energy, how can any organization which provides so much aid to the U.S. military be considered pro-ecological? (And even if the U.S. military were to drastically cut its emissions and become “green,” I would not wish to see “sustainable” invasion, occupation, and drone bombing.)

Blockchain technologies

Cryptocurrencies are but one of many applications of blockchain technology. Bitcoin is the best known of these technologies, which are chiefly distinguished by maintenance of global consensus about the state of a shared ledger without a central authority, primarily because of its formative historical role and significant financial value. In other aspects, it performs weakly: fees for sending bitcoin are frequently high, and transactions may take hours to be confirmed, in addition to its previously discussed extreme energy inefficiency.

It should not be surprising that the first iteration of a still-early technology has proved to perform poorly on a number of important axes. Less than ten years after Bitcoin was introduced, programmers have already developed more generally useful blockchain software. For example: Filecoin is designed so that, rather than “miners” performing a lot of useless computations as with Bitcoin, network participants receive rewards for allowing others to utilize their unused storage space to store data in the “InterPlanetary File System.”

Another interesting project to briefly discuss is Ethereum. Unlike earlier blockchain technologies that provided users only a limited set of predefined operations, Ethereum exposes its users to a virtual machine capable of carrying out any computation. Hence it is possible to build arbitrary applications on top of Ethereum’s decentralized infrastructure, similar to the way many organizations have built applications on top of the centralized infrastructure of companies like Amazon or Google. A fundamental aspect of the design of the Ethereum Virtual Machine is “gas”, which represents the cost paid by the person who wants some code run to the person who is using their computer’s resources to execute the code. Even if the program fails to achieve its intended goal, the user must still pay gas for the resources they used, giving developers strong reasons to carefully consider the resources their programs consume. I have not attempted to create one of these applications myself to give a first-hand assessment of the developer experience, but I believe a mechanism like this could be helpful in forcing programmers to consider the ecological implications of their actions. And in part due to environmental concerns, researchers are investigating the possibility for Ethereum to use a “proof of stake” mechanism for validating transactions, which could be far more energy-efficient than the “proof of work” system it currently uses, much like Bitcoin’s.

Despite these and other hopeful projects, blockchain technologies so far remain an intensely resource consuming sector, and I consider them of net negative social value for the global majority. An especially ugly example is on display in Puerto Rico following the devastation of Hurricane Maria, where capitalist investors began buying up property and monopolizing the limited electrical resources there to build a practically tax-free “crypto-utopia” for themselves, in the latest iteration of colonialism for the long-suffering residents of that exploited island.

Besides public blockchains like those discussed above, private blockchains are also under development for coordination between distinct entities throughout a supply chain. The vice president of food safety for the infamously anti-worker Walmart bizarrely claims that a food traceability pilot project will usher in an era of unprecedented transparency for customers by sharing knowledge of the labor conditions of its agricultural workers. However, it seems clear that labor unions and solidarity groups could use this previously obscured information about the relationships between workers and consumers against corporations. Promises of radical transparency notwithstanding, the information contained within private blockchains will likely remain private to those with a capitalist-defined need to view it.

Other decentralizing technologies

Despite the hype pervading many discussions of blockchain technologies, they are far from being the universally applicable software engineering approach to gain the benefits of decentralization, such as immunity to a single point of failure. Critical software already in use for years before Bitcoin existed and still in wide use today includes peer-to-peer file sharing protocols such as BitTorrent and distributed version control systems such as Git. Research, design and development work is ongoing on these technologies as well as newer non-blockchain distributed systems that do not appear to share the same drawbacks in energy inefficiency as blockchain technologies.

One such project is dat, a peer-to-peer protocol for securely exchanging archives of files over networks either public (e.g., the internet) or private (e.g., a sneakernet). An interesting and useful application of this building block is the experimental peer-to-peer Beaker Browser, with features including: the built-in ability to host content directly from the browser; private file sharing involving no intermediary such as Dropbox; and future-proof, unbreakable links thanks to content addressing. The peer-to-peer approach of downloading a file from anyone who has it—preferably a peer nearby—has the potential to create significant ecological efficiencies compared to the current model of many clients downloading from a smaller number of centralized servers.

Finally, movements towards decentralization of physical manufacturing via hardware such as 3D printers show ecological promise if properly designed and deployed—though I consider that under capitalism, the probability of such care is remote. Potential benefits include localization of production, thereby reducing emissions associated with transport, and the fact that “additive manufacturing” does not necessarily create scraps in the way that traditional “subtractive” manufacturing does. There are many other possible benefits and drawbacks, which I cannot discuss here; for some existing work in this area, see the 3D Additivist Cookbook.

With this, we conclude our whirlwind tour of the state of the art in ecological computing, and zoom back out to the guiding question posed by the title of this article.

Prospects

Prospects for computing, ecological or otherwise, are inseparable from prospects for society generally. It is depressingly easy for me to imagine a future much like the oppressive present, only with more advanced technology than ever in the hands of the ruling elite and their (private or “public”) security forces and with some of the most disastrous possible changes in climate having come to pass. Yet, depending on the outcomes of various social struggles, a very different world is possible.

Pessimism

Frantz Fanon said: “Each generation must, out of relative obscurity, discover its mission, fulfill it, or betray it.” In contemplating what is that mission today, I perceive clear parallels between the position of relatively privileged tech workers and the petty bourgeoisie in post-colonial nations, as described by revolutionary leaders such as Amílcar Cabral in his address “The Weapon of Theory” to the 1966 Tricontinental Conference of revolutionaries from Africa, Asia and Latin America. In this speech, Cabral discusses the vital necessity for anti-colonial movements to include an analysis of imperialism, nationality and class in order to attain true liberation:

In order not to betray these objectives [of national liberation] the petty bourgeoisie has only one choice: to strengthen its revolutionary consciousness, to reject the temptations of becoming more bourgeois and the natural concerns of its class mentality, to identify itself with the working classes and not to oppose the normal development of the process of revolution. This means that in order to truly fulfill the role in the national liberation struggle, the revolutionary petty bourgeoisie must be capable of committing suicide as a class in order to be reborn as revolutionary workers, completely identified with the deepest aspirations of the people to which they belong.

(NOTE: Cabral’s use of this phrase is intended metaphorically rather than literally, in line with similar usages from the same era such as that of Huey Newton of the Black Panthers, whose autobiography was titled “Revolutionary Suicide”. Today I would prefer a term such as class betrayal over class suicide.)

This path to liberation is diametrically opposed to the path of “diversity” and assimilation into oppressive structures paved by Black faces for white supremacy such as Barack Obama. Every oppressed group contains numerous such heroes and misleaders (or “compradors”) alike.

With some understanding of how our own work fits into this larger context, it is obvious that as technologists, we certainly cannot “save the world” with clever hacks. Yet without a critical mass of sympathetic and participating technologists, it will be far more difficult for anti-capitalist social movements to counteract high tech surveillance and repression and to take advantage of the novel communications and organizing possibilities enabled by modern technology. Between the scale of the resources at the disposal of state security and “surveillance capitalist” corporations and the racialized and gendered composition of the high tech workforce, these organizations have little difficulty in employing conformist and/or actively reactionary tech workers to develop and refine techniques for the manipulation and repression of communities in struggle. The distorted perspective of this privileged class is expressed in the ugliness of the biases and oppressive applications evident in today’s machine learning technology.

The division of labor in the computing industry must be situated in the greater history of capitalism’s use of nation, race and gender to divide workers and organize productive labor, including reproductive labor and care work. The contrast between the workers who physically produce computers and the white and male domination of the software industry in nations like the U.S. could hardly be more stark. Although both groups are indeed workers, the Global South proletarians who predominantly assemble computers have a far more antagonistic relationship to imperialism compared to White settlers and the “diverse” others aspiring to a similar status, who largely benefit from and support imperialism. From this perspective, proposed solutions such as nationalization of industry or natural resources have hugely varying significance and consequences across geographies and sovereignties. For example, I support collective control over resources in places like the Democratic Republic of the Congo, and I oppose the U.S. settler state taking even more direct control over technology than it exercises now.

However, I do not know how to destroy the surveillance apparatus and accumulated data of imperialist security agencies, nor how to achieve social control over the computing infrastructure of a company like Google, nor how to alter the physical workings of computing supply chains and product life cycles to achieve workers’ flourishing and ecological sustainability, nor how to reconcile broader questions of decolonization, Black self-determination, and socialist democracy, particularly in settings where colonizers comprise a numerical majority. What I do know is that none of these things will happen without a fight.

It would be utterly self-defeating to meekly allow the ruling class to continue pillaging the Earth and oppressing humanity in the name of “nonviolence,” as liberals will endlessly allow with their attempts to police the legitimate means of resistance to the limits authorized by imperialist states. For if capitalists remain free to continue acting as they have been, we can expect to observe far more of the same violence we observe now: more environmental destruction; more displacement; more war; more ableism, racism, patriarchy and oppression of all kinds, rising to the level of genocide for “disposable” populations whose labor is not essential to capitalism. I agree with the assessment of Cooperation Jackson and the Community Production Cooperative:

…if these technologies [of the “Third Digital Revolution”] remain the exclusive property of the capitalist class and the transnational corporations they control…the end result will be a global dystopia, a social nightmare predicated on massive poverty, lawlessness, state repression, and ever greater human disposability.

Some live in this dystopia now. As William Gibson said: “the future is already here—it’s just not evenly distributed.”

Optimism

The above scenarios are not inevitable.

While sobered by these warnings and realities, I am also inspired by the positive efforts of (among others) the above-cited Cooperation Jackson and the Center for Community Production in Mississippi to counteract class, racial, and gender divides in the world of high-tech, both in terms of workforce composition and in terms of who controls and benefits from the technology. These initiatives point towards a model of public/community partnerships between municipal governments and collective institutions such as credit unions and worker and consumer cooperatives to enable community control over modern technology such as manufacturing equipment, producing directly for community consumption rather than a capitalist supply chain, while respecting ecological constraints. I urge those with the spare funds to donate to support their attempt to take a portion of this technology’s development away from the entities critiqued above and put it into the hands of their community. They also call for similar projects in other communities, which I agree are sorely needed.

Socialism can only be built from such deliberate, conscious efforts. However humble are the early stages of this lengthy and uneven process, in the final analysis, the scope of our work must match the colossal task before us. When examining the fractal pattern of today’s immense inequality—at the largest scale between the Global North and the Global South, but also within nations, regions, municipalities, and households—it is clear that massive reparations are absolutely necessary to redress the material disparities created by centuries of plunder, as well as future dislocations already nearly certain to be caused by climate change. Questions of who receives the reparations, what form they take, and how they are subsequently administered are inseparable from questions of how to organize coalitions powerful enough to win them. Obviously, oppressor classes will not willingly part with the fruits of their conquests; as C.L.R. James wrote in “The Black Jacobins”: “the rich are defeated only when running for their lives.” Diverse methods of organization will be required, and every means of struggle should be employed at its appropriate time. One such is critical appraisal, including of the tools at our disposal to do this work.

Nearly fifty years after it was written, Amiri Baraka’s poetic critique of technology still resonates with me after lonely years of alienated labor on a laptop:

A typewriter?–why shd it only make use of the tips of the fingers as contact points of flowing multi directional creativity. If I invented a word placing machine, an “expression-scriber,”*…*even the final xpressed thought/feeling wd not be merely word or sheet, but itself, the xpression, three dimensional–able to be touched, or tasted or felt, or entered, or heard or carried like a speaking singing constantly communicating charm. A typewriter is corny!!

The so called fine artist realizes, those of us who have freed ourselves, that our creations need not emulate the white man’s, but it is time the engineers, architects, chemists, electronics craftsmen, ie film too, radio, sound, &c., that learning western technology must not be the end of our understanding of the particular discipline we’re involved in.

Baraka’s meditation on what technology could be points towards a type of diversity and inclusion I can support: not a superficial change in the complexion of oppressors and oppressed, but a fundamental change in who participates in decision-making, yielding radical differences in what technology is designed, constructed, and used. Such a change in the character of computing can and must include ecological sustainability as a core requirement for the field’s existence. For despite the depth of the influence of imperialism in the history of their development, computers need not work in the often destructive way they do today. Neither does technical education need to be synonymous with “cultural invasion” stulting the imaginative potential of the colonized people supposedly being educated. We can build social computing spaces in which people of diverse abilities, ages, nationalities, and genders can freely explore their common and unique interests in settings designed to encourage collaboration, artistic exploration, and natural physical interaction for people inhabiting all types of bodies. And in tandem with coming together and changing the way we relate within our own localities, we can build international bonds consciously designed to reverse the ordinary exploitative relationships between imperialist core and periphery, putting what resources we can at the disposal of oppressed peoples, including the Native peoples of the land we live on.

There are immense difficulties to be faced in the course of these attempts to widen the sphere of who controls technology, just as there are in every effort to build solidarity among groups facing differing types of oppression and participating in each other’s oppression. Yet not even the most brutal and total methods of control are infallible, as my formerly enslaved ancestors proved in their eventually victorious struggles to free themselves from bondage. To demonstrate our respect and gratitude for all that our forebears did for us who were not-yet, we must show the same responsibility towards future generations.

A basic aspect of this responsibility is to ensure that they inherit an Earth fit for their habitation. This can only mean organizing for the seizure of the productive forces and wealth currently monopolized and misused by a selfish few, in every field of endeavor including computing, and the redirection of these social resources towards the development of the potentialities of the many and the healing of our mother Earth. These are no simple tasks. Nevertheless, in the words of Assata Shakur, it is our duty to fight and our duty to win.

“If you dare to struggle, you dare to win. If you dare not to struggle then god damn it, you don’t deserve to win. Let me say peace to you, if you’re willing to fight for it.” - Fred Hampton