Tagged AI history · show all articles
Meta did not set out to start an open-weight movement. A gated research release leaked within a week, and everything that followed — the commercial license, the benchmark claims, a rival's answer — was improvised in response to that accident.
Long before deep learning existed, a mathematician warned that a machine pursuing a fixed purpose could not be trusted to want what its operator meant. The distance from that warning to a shipped alignment pipeline is the field's actual, dated history.
Four calling conventions, each dated to a real announcement: a manifest borrowed from web APIs, a schema built into the model call itself, incompatible framework wrappers, and a standard now governed by the vendors who once competed on incompatibility.
Grading a model once meant scoring a string against a reference answer. Grading an agent required borrowing a games lab's tools, discarding them, and rebuilding evaluation from scratch, in a documented, dated sequence of its own.
Every era of AI agent architecture answers one question differently: what decides the next action. A 1971 planner, a learned reward signal, a prompted reasoning loop, a delegated sub-agent, and a shipped SDK are five dated, sourced answers.
Anthropic's own dated record — a 2021 founding, a Constitutional AI paper that preceded its first product, a Responsible Scaling Policy that has since restricted a real model's deployment — is more precise than the branding built on top of it.