Explainers
The machinery, explained.
Four pieces on how modern AI systems are built, served and read. No equations required.
How a transformer actually works
Tokens, attention and the stack — without the math.
Read → 02Training vs inference: where the money goes
Why serving a model can cost more than building it.
Read → 03Small models on ordinary hardware
Quantization, memory and what actually fits on your desk.
Read → 04How to read an AI paper without drowning
A practical order for abstracts, figures and limitations.
Read →