Points to make: BCS observation

Phase 1: 1960--1970s: AI -> CogSci trajectory from basing AI on what's
understood in Psych/Ling/Anthro/Socio etc. and developing
computational implementable approaches to understanding e.g. human
language.  HMMs

Phase 2, Late 1970s-early 20th: Evidence-based, but with relatively
little evidence by today's standards: a few millions of words of
written (American) English, a few 10s of hours of spoken, simple
probabilistic models, _no_ reference to human process.  ANNs.

Phase 3, 2010s to 2023:  Step-change in the size (of data, of models,
of compute power) and in the complexity of the architecture (LLMs) and
the underlying mathematics.

Apothesis: Release of ChatGPT in April 2023 (?).

But one thing hasn't changed: it's all about the _form_ of the data,
not the substance (BCS).