Transformers

Attention-based architectures radically expand the scale and capability of language models.

Why this milestone matters

This entry marks a change in the relationship between human intent and executable systems. It is presented as a navigational marker inside a wider, overlapping history—not as proof that a single date, person or institution acted alone.

Evidence

Evidence type: primary research.

Attention Is All You Need ↗

Editorial note

CodeHistory keeps timeline priority claims revisable. Where the surviving record is retrospective, that limitation is stated rather than silently converted into certainty.