From Attention to Agents

Where does the optimization happen - in the weights, or in the context?

Twelve chapters bridging the late-2019 CNN era to the September-2026 reasoning-agent frontier. Each chapter turns one lever - architecture, scale, pretraining efficiency, alignment, democratization, context, test-time compute, RL at scale, inference systems, agency, measurement - and the last chapter stacks them. Written for readers who knew CNNs in 2019 and watched attention just enter NLP.