Reading guides for AI research
Each guide starts from a question people ask about AI results, links the papers and lessons that address it, and says where the evidence stops.
What does a benchmark score let you conclude?
How to read an AI benchmark · 3 sources
Start with the task, the evaluation rule and the resources used. A final score describes a tested setting; it does not establish that two systems were compared under the same conditions.
Which bottleneck does a change actually reduce?
Where AI inference spends memory and time · 4 sources
Keeping fewer tokens, reading less cache and generating several tokens at once change different parts of inference. Identify the measured resource before comparing a memory reduction with a speed improvement.
Does agreement between judges establish validity?
When to trust an LLM judge · 4 sources
Agreement between evaluators, agreement with experts and sensitivity to a prompt are different measurements. Name the property you need before deciding whether a judge’s score answers your question.