Post-apocalyptic software engineers
Recently I have started wondering, if I somehow survive the apocalypse and become a member of a few hundred Post-apocalyptic human society, what would I do? How I would be useful in this new barebone community? Then I realised there are many variables in this topic to consider.
First of all, probably COVID-19 made me think about this sort of future. At the time of writing this, the only "apocalyptic danger" I had seen before COVID-19 was the 2012 Mayan conspiracies. No one died due to the Mayan calendar, but this time things got very real, very quickly. Almost the population as big as Hong Kong have died. But what would happen if the number of alive people was equal to Hong Kong's population? Or even less, a couple of neighbourhoods only stayed alive? And if I was one of that "lucky" population as a software engineer, would I be of any use at all?
If the population is rapidly dying due to some massive-scale virus, that population would still have some time to transfer some knowledge and ownership of things. The whole world would be in chaos, but I expect at least a dozen engineers to come together as doomsday committees to save the last bit of information lying around. What I mean is, that in this disaster scenario, we still would like to save the existing digital knowledge of humanity. I can imagine teams coming together and moving common and important data to some servers physically close to each other as efficiently as possible. Those servers should also be close to some sort of cheap power source. That can be as big as a solar/wind power plant or just a couple of generators that run on diesel. Once a certain power source is agreed upon, in terms of the server side, most of our servers would be cold storage options. Our aim here is not to access our data fast, but rather to make sure we can access it after many years.
We don’t have much time looks like
In that case, the best thing to do is probably to open your servers to the internet and hope that someone finds it out and accesses your data to carry humanity forward. Over many years, we have built systems that prevent unauthorised access, but now it is time to reverse it. Share your server's IP address, and create a root user with a pre-agreed password. The doomsday committee can agree that everyone should reset their passwords for useful data to asdf12345 so we can all access the leftovers once they are gone. Oh, one more thing. Make sure to change your distributed systems, servers, and NAS to the most redundancy mode. So once you are away, the system can survive as long as possible with as many hardware device failures as possible.
Does anyone even care about my data?
They may. Probably for the majority of us, it won't matter. A code for an image-sharing social network or a food delivery app may not be necessary for Dooms Day, but a code for a global file indexer might be useful to access everything else that is out there lying on intranets.
It happened, what now?
I am thinking about how I can be useful immediately in the doomsday. I can't harvest crops, I can't treat people, I can't teach kids, I can't organise/lead people. But I can help them to access the knowledge. I can help farmers, doctors, and architects to access the knowledge. Assuming that we can sort out some cheap electricity by using modular generators we have left over and assuming the doomsday committee shared details of how to access Wikipedia, AWS, and Google servers physically, I can probably go build a crawler that finds all the knowledge that we would need to rebuild the society. We could just access the British Library and try to use actual books for the knowledge, but even the British Library uses technology to catalogue all the knowledge for the whereabouts of those books in their vast collection. As a software engineer, it would be useful to revive that catalogue on a backup power, for example, to recover people's access to knowledge.
2025
Originally I started writing this article in 2020 during Covid, which made me reconsider my position in society for various reasons. Since then I never completed this article because other things in life happened. Now in 2025, I think the most important piece of "digital knowledge" that I would prioritise as an apocalyptic engineer would be backing up large language models. Essentially, in the past few years, as humanity, we have managed to compress the world's knowledge into a single file that takes a couple of gigabytes of space. This is just incredible. I know LLMs can hallucinate and the information that they spit out needs to be double-checked, but in the post-apocalyptic world, even a hallucinating friend will be useful. They would be the very first source of knowledge for me to back up and utilise considering how much value I can gain with those couple of gigabytes vs any other piece of knowledge that I can fit into the same amount of space.