Twenty papers that built the software industry, five from each of four fields, each explained on its own page from the problem outward: what the world looked like before, what broke, what the authors proposed, and how it works, in plain words, with diagrams and an interactive simulation you can play with. Every page links to the original paper and to the best free explanations elsewhere. Read a category in order, or jump to the one you keep hearing about.
time · consensus · scale
Five papers that define how many machines act as one: what "before" means without a shared clock, how a thousand computers run one program, how a store stays up when parts of it are down, how a cluster agrees, and how a database spans the planet.
tables · logs · trees
Five papers under every database you have used: why data became tables, how a crash never loses a committed write, why modern stores write sequentially, how Google kept a sorted map across thousands of machines, and why the log became the centre of data infrastructure.
files · traces · tails · schedulers · sandboxes
Five papers from inside the machines the cloud is made of: a file system that expects disks to die, a way to see one request cross a hundred services, why the slowest one per cent decides your latency, the scheduler Kubernetes came from, and the tiny virtual machine under serverless.
depth · vectors · attention · scale
Five papers that explain why the software you use now talks: the network that started the deep-learning era, the trick that turned words into geometry, the architecture under every large language model, the moment models learned tasks from a prompt, and the laws that told labs to keep scaling.