Reading

Vaswani et al., NeurIPS 2017

The architecture is elegant but the training stability tricks in the appendix are what made it work in practice.

DeCandia et al., SOSP 2007

Read this to understand eventual consistency — not the theory, the engineering tradeoffs.

Ken Thompson, Turing Award Lecture 1984

Four pages. Changed how I think about software supply chains permanently.

Ongaro & Ousterhout, USENIX ATC 2014

If you've only read the Paxos papers, read this. Same guarantees, actually implementable.

Dean & Ghemawat, OSDI 2004

Paul Graham

grugbrain.dev

Complexity is the enemy. Return to this every six months.

John Ousterhout

Better than Clean Code for experienced engineers. Deep modules, not shallow ones.

Richard Hamming

Chapter on "You and Your Research" is worth the whole book.

Richard Feynman

Reversible computing and the thermodynamics of information. Underhyped.