Reading systems papers without drowning

You do not need to read every paper cover to cover. You need a way to decide which papers deserve a deep pass.

First pass

Skim:

• Abstract
• Introduction
• Conclusion
• Figures

Ask one question: What claim would change how I build systems if it were true?

Second pass

Only continue if the claim matters. Then read the evaluation carefully. Look for assumptions that do not transfer to your environment.

Capture

Write a short note with:

• the problem framing
• the key mechanism
• what you distrust
• one experiment you could run

That is enough to make research reading cumulative instead of ephemeral.