Welcome to the new pandemic - anxiety. COVID-19, geopolitical tensions, and environmental calamities have taken their toll. Societal collapse is a concern for many people, and over a third of Americans believe that they're living in the end times.
If they're right, and society changes radically in the next few decades, what would happen to computing? Some technology experts are working to ensure that it wouldn't completely disappear.
To protect computing during a societal collapse, we must understand what it would look like. Scientist Joseph Tainter defined it as a reduction in socio-economic complexity. Society becomes more complex and global, bringing huge benefits, but eventually the complexity becomes difficult to maintain, and it devolves.
Events like climate change, growing distrust in governments, and war strain the system, making this more likely. A growing community of experts worry that this is happening now. KPMG recently re-assessed a seventies-era MIT study predicting social collapse by the middle of this century, and found it to be on track.
A whimper, not a bang
We don't have to wait for collapse to begin, warns. Bill Tomlinson, a professor at the University of California, Irvine specialising in ICT and sustainability. He appropriates a famous William Gibson quote:
The collapse is already here. It's just not evenly distributed.
He cites climate migration, which experts predict will grow, as evidence of a collapse already underway.
While scenarios like nuclear war or virulent disease could level civilisation overnight, Tomlinson foresees a slower decay:
It feels like it'll be a gradual and bumpy road down. There will be lots of gradual declines, where the brownouts will be more frequent, and various resources won't be as available.
There's a continuum between a national toilet paper shortage and a Mad Max-style feudalism that we'll hopefully never see. How we act now will affect where we land along it. Somewhere in there lies a world in which we've failed to make a seamless transition from abundant fossil fuels and where energy is harder to come by. In that scenario, technology supply chains have withered, and we've lost our ability to build the kind of tech we're used to. Lifestyles and technologies might resemble those of the early 1900s.
How today's tech could fail us tomorrow
If society does reach that point, ICT will be more important than ever, says Tomlinson:
We may not be up to the task as a group to address this kind of problem. On the other hand, the thing we have that allows us to deal with broad scales of time and space is technology.
When used properly, IT helps us to work faster and retain more. It helps us become more than we are. In a society that has retracted in reach and expertise, that would be more important than ever.
Students of collapse worry that our current technology might not be well suited to the job. They see challenges running all the way up the tech stack, starting with an electrical grid that might no longer be as reliable. Technology would need to be more resilient to intermittent failure.
A lack of natural resources and energy could also affect the cloud computing infrastructure that we've spent the last decade persuading everyone to embrace. Data center operators are exploring free-air rather than evaporative cooling, but these are still in the minority. There are thousands of water- and energy-guzzling facilities in operation today. This makes Richard Heinberg, senior fellow at the Post-Carbon Institute, nervous. He says:
I'm so fearful about the degree to which we rely on the cloud for all of this vital data. Redundancy is a key aspect of resilience. You make things more redundant by having local copies available in more places. So we need alternatives to the cloud.
Cloud infrastructures and the software that runs them also have another characteristic that is a liability in an era of collapse: complexity. Everything from fourth-generation languages through to infrastructure as code are abstractions that hide a rich sediment of work stretching back years. That's a problem for tech resilience, warns software engineer Rett Berg, founder of CivBoot, a nascent project to provide a blueprint for rebooting technology production during collapse:
We're building abstract layers on abstract layers on abstract layers, and it lets us build really cool stuff, but eventually, it's going to fall over.
Building a resilient tech stack
Berg is among a small, loosely-coupled group of people racing to create a simpler, more resilient tech stack. This peri-collapse tech stack begins with sustainable hardware. as modern computers begin to fail, some fret that cutting-edge processors will become scarce. In any case, when people become more interested in managing greenhouse temperatures than mining crypto, the emphasis will be on simplicity, compactness, and energy efficiency.
These requirements call for new approaches, says Quebec-based sysadmin Virgil Dupras, creator of the CollapseOS operating system:
If I was in that situation, the first thing I would reach for is an old machine to scavenge for. Old computers have chips that are really easy to desolder.
CollapseOS is accordingly old-school. It is a complete package for developing, assembling, and running programs on Z80, 8086, 6809 and 6502 chips. After originally writing it in assembly language, Dupras switched to Forth because programs written in it are extremely compact and require only minimal assembly.
Berg's CivBoot was inspired by CollapseOS but takes an alternative approach. It is a blueprint for a production facility that can create all the technology necessary to replicate itself. A facility in one community could churn out the software and tooling to create others like it, boot strapping technology production locally across the country. That needs more powerful, readily available processors, produced domestically, says Berg:
To make a system self-replicating and understandable, you need other capabilities, such as the ability to run CAD tools.
Semiconductors will be too complex for a CivBoot facility to make. Instead, he sees thin-film transistors as the way forward. This uses chemical processes to deposit purified materials on a glass substrate. It's still a complex thing to make, requiring university-level funding and research. In the meantime, he is concentrating on the other aspects of CivBoot, including all the tools necessary to fabricate the CPU such as CAD systems and production control.
Storage and communications
Aside from computing hardware and software, society will need at least two other resilient components: information, and the ability to store and send it.
Technologists have turned to decentralisation as a way to store information resiliently. The Interplanetary File System (IPFS) is a project to distribute web sites and other files among thousands of participating computers online, creating a vast, distributed cache of multiple copies. People are already using it to create 'shadow libraries' of academic and other literature, preserving it even if centralised sites go down.
The IPFS could help to distribute that information more reliably if the internet itself becomes less usable, says Tomlinson:
It happens to deal well with intermittency, which is likely to be one of the characteristics of a collapsing technological infrastructure.
Will the internet go away? Tomlinson imagines pockets of it that will operate close to normal as collapse continues. Nevertheless, aside from brownouts there is also the spectre of a 'splinternet' - a Balkanised internet with factions cordoned off by nation states or even local authorities.
Don't Look Up
These tech proposals are resilient, which is one of the biggest requirements before, during, and after collapse. The biggest problem is getting people to adopt them. Most people still prefer inscrutable formats like docx and xlsx because of the extra functionality that they support, even though human-readable text files might serve us better in an age of collapse. And while thousands have adopted IPFS, it is still a niche service. People simply want things that are easy to use, Tomlinson points out:
Industrial civilisation tends to push efficiency, even if that causes a bit of brittleness. That efficiency tends to crowd out these alternative ways that you might structure technology.
The community of people working on technology for collapse is small, complains Dupras. Mirroring the movie "Don't Look Up", few people want to accept the concept at all, turning instead to technological optimism and the idea that we'll reach escape velocity with tools like AI. But he says we should at least plan for an alternative. He sums it up:
Nobody's peeking into the abyss.
That isn't slowing him down. His latest project, DuskOS, is a simple operating system designed for what he sees as the first stage of tech collapse, where supplies of new computers are slim but modern desktops and laptops still work. It will allow people to manipulate ailing machines more simply, keeping them operational as components fail.
Collapse isn't a topic that most people enjoy thinking about. However, faced with increasingly grim news headlines and scientific data, Heinberg recommends a mindset that he calls defensive pessimism. In short, plan for the worst. With that in mind, resilient tech could use a lot more work, and funding, to move it into the mainstream.