Reading List

Sagan on why a society that stops investing in curiosity-driven research cannot sustain itself.

A 1966 address to science teachers. On the difference between knowing the name of something and actually understanding it.

The gene-centered view of evolution. Rewired how I think about natural selection, cooperation, and why organisms exist at all.

The most important pattern in AI research: methods that leverage scale always win over methods that encode human knowledge. Short and still being ignored.

Which problems are most important to work on? Rigorous analysis of AI safety, biosecurity, and other pressing risks.

Research on catastrophic and existential risk governance and decision-making under deep uncertainty.

Pandemic preparedness, biosecurity research, and global health security policy.

Policy and research working to reduce nuclear and biological threats globally.

A framework for thinking about how technological progress could produce catastrophic outcomes even without any actor intending harm.

Architecture case studies of real high-traffic systems. How Discord, Stripe, and others actually work.

Thompson's 1984 Turing Award lecture. Why you cannot fully verify the integrity of a system you did not build yourself.

The 1978 paper that introduced logical clocks. Still the clearest explanation of why distributed time is hard.

The paper behind much of how we think about eventual consistency and availability tradeoffs in distributed databases.

Software-defined memory architecture that disaggregates DRAM from compute nodes. Relevant to how memory-intensive workloads can be scaled without physical hardware constraints.