Common Fallacies

This page is work in progress.

Carbon Tunnel Vision

https://environment-analyst.com/images/Carbon-tunnel-Vision.jpg

Here is the thing: Yes, the climate crisis requires us to reduce carbon emissions. Fast. But what created the emissions in the first place, was not just fossil-fuels but the unsustainable human & economic development strategy we were on (e.g. ‘through economic growth, we can lift everyone out of poverty’).

Second, yes global warming is caused by high concentrations of Green House Gas Emissions (GHG Emissions), but there are many second-order effects which are negatively impacting our planet, our soils, forests, biosphere, etc. which make it more increasingly difficult to reduce GHG Emissions to begin with.

A simple example are the oceans, which are the largest ‘carbon sinks’ on our world through their biosphere such as algae. As we start to do deep-sea mining to fuel our hunger for cobalt, lithium and other rare-earths and metals, we increase the risk of damaging that biosphere, which will in turn mean that the oceans lose their carbon-storage capacity. On our quest to fighting the climate crisis, we have to make sure we don’t damage the storage & absorbing capabilities our planet already has. All the current calculations on the reduction path are build upon on the assumption that those storages will remain in tact.

So when we think about the impact that software has, indirectly through the infrastructure, hardware, and energy it requires to run, we can’t just think about GHG Emissions as a proxy for environmental impact - we really have to work on the environmental impact as a holistic set of metrics. Here is an excerpt from the United Nations, Life Cycle Initiative which is working on a globally unified set of environmental impact indicators.

  • Greenhouse gas emissions and climate change impacts
  • Health impacts of fine particulate matter
  • Water use related impacts – water scarcity and human health impacts
  • Land use related impacts on biodiversity
  • Acidification and eutrophication, human toxicity
  • Natural resources – mineral primary resources
  • Land use impacts on soil quality
  • Ecotoxicity

Now if you are thinking ‘must of these having nothing to do with software’. You are right. But now think of the factories which make chips and the chemicals which are used for that. Where do those go? How do all the rare-earths, the lithium, cobalt, neodymium and many other rare-materials get mined? What about data centers, do they require land to be closed with cement? The diesel generators which run weekly to monthly for testing them? The water in the cooling systems that has chemicals added to avoid corrosion?

https://euinfo.rs/plac3/wp-content/uploads/2021/04/eotpad1.jpg

https://scx2.b-cdn.net/gfx/news/hires/2020/5ef48bcd38874.jpg

Yes software itself has none of these issues. But software doesn’t physically exist, it only exists because there are computers - made from hundreds of minerals & metals - turn energy into data processing, transport and storage capabilities that the software can use. If you unplug a computer, the software it was running ‘seized from existence’.

Your software is driving the demand for more hardware through the amount of processing, transport and storage capacity it requires. Minimizing resource usage and thus making an impact across all those metrics is your responsibility. Alongside with telling suppliers of hardware and infrastructure to do better, fast.

Green Energy

More Software in Green

Digital is always net-positive

A computer science professor once proclaimed “Any analog process turned into a digital process is by nature more efficient.” Of course he meant to say ‘energy efficient’. Our obsession with efficiency often does not include the total environmental (or human impact, labor effort needed) costs of the thing we are transforming.

It’s often too complex for the human mind. But isn’t that what we have computers for? Shouldn’t we be able to quantify the total costs & impacts of transforming a process, product from analog to digital?

In either case, anyone who has build software or hardware products, likely already knows that in many cases digitalizing a process did not lead to ‘efficiency gains by nature’. It has become a go-to-solution for many areas of our society, especially because it neither negative or positive effects are often not verified anyhow.

Maybe its a believe worth questioning.