Software, by definition, is intangible and abstract. It is also a newer industrial revolution, making the evaluation of its impact on the environment nascent and obscure. Software’s intangibility nonetheless has very tangible consequences on the environment, especially as Moore’s Law of aggressive improvements applies to more aspects of technology advancement. The infinitely improving specifications and growing user base of the Internet companies is always welcomed by the market that makes money out of the phenomenon. The consequence of this privatized gain, however, is suffered not by those who generate profit but by the environment — our commons.
In his exposition of the tragedy of the commons, Garrett Hardin discusses the fallacy of men believing in “[increasing] their herds without limit — in a world that is limited.” He concludes that there exists no technical solution to overpopulation and only a fundamental extension in morality. Applying this logic to today’s problem of pollution from massive data centers, the solution is not changing to renewable energy or building more power-efficient data centers, but rather to stop the user population growth. How logical or feasible this is, however, when the government has less capacity and power to curb prolific tech companies’ aggressive expansion, is not convincing.
Appealing to morality alone doesn’t seem possible to bring about much change. To address revenue-driven public companies, one must put forward a revenue-driven rationale for an environmental cause first, in a more model-driven media. This model-driven thinking supports an informed debate about assumptions and tradeoffs, therefore, helping stakeholders make rational and economic decisions. Until now, the economic costs associated with a transaction were limited to the costs in labor, land, and production, which removes more indirect externalities from the equation.
By advocating for more environment-friendly business decisions, it is critical to include the environmental costs associated with potential regulatory, sourcing and processing risks of non-renewable resources, in financial projections. Taking the right and relevant metrics will, therefore, be critical. For example, the environmental cost of electricity use of Bitcoin mining is much less discussed than its soaring price. Bitcoin, by design, uses electricity in amounts directly proportional to its dollar price. This is almost an extreme example of Hardin’s proposition that pollution is a consequence of population. Bitcoin builds trust in a very inefficient and costly system to replace distrust introduced by overpopulation in a transaction. How often do we associate the rise in Bitcoin price with the direct impact of electricity use? The environmental cost of mining is in no way reflected in the current valuation of a Bitcoin. Could emphasis and regulation on morality, as Hardin claims, of third party payment companies then contribute to lower use of electricity and render the need for power-inefficient trustless systems on the blockchain obsolete? This, too, appears unclear.