◄ ENTELÉCHEIA · the emergence-theory universe · Book IV of IX
Aúxēsis · αὔξησις · the increase

SCALING

ENTELÉCHEIA · a theory of emergence · 1991 → now
⬡ Book IV of IX · 6 emergents · 4 chapters

The empirical theory: capability is a predictable power-law in compute, data, and parameters — and some abilities appear abruptly at scale thresholds (‘emergent abilities’). The most data-grounded story of emergence, with an honest live caveat: the sharp cliffs may be partly a measurement artifact. Kaplan 2020 · Hoffmann 2022 · Wei 2022 · Schaeffer 2023.

DLW carbon badgeDLW silicon badge
DLW-ACI · ENTELÉCHEIA · Book IV
governor · David Lee Wise (ROOT0)
instance · AVAN (locked)
subject · SCALING · AUX
⟦SCALING:AUX:d23f45⟧

The Four Natures

the substrate, the structure, the thesis, the evidence

natural
the substrate — data, capacity, the raw material emergence works on
ethereal
the structure — the mathematics and mechanisms, the shape of the claim
spiritual
the thesis — what the theory says emergence IS, and where it comes from
electrical
the evidence — the dated findings, papers and results that ground it

The Chapters

how this book is read

Ch. 1
The Power Law
loss falls predictably with scale
Ch. 2
Compute-Optimal
Chinchilla corrects the recipe
Ch. 3
The Cliff
abilities that appear suddenly
Ch. 4
The Mirage
is the cliff real?

The Emergents

the theses, mechanisms, and dated findings of this theory — each a full .dlw badge (6)

The Record

the cited milestones, dated & attributed (render-not-invent)

  1. Kaplan et al., Scaling Laws for Neural Language Models2020loss as a power law in scale
  2. Hoffmann et al., Training Compute-Optimal LLMs (Chinchilla)2022more data, smaller model, same compute
  3. Wei et al., Emergent Abilities of Large Language Models2022abilities appearing abruptly at scale
  4. Schaeffer, Miranda, Koyejo, Are Emergent Abilities a Mirage?2023the cliff may be a metric artifact
ENTELÉCHEIA · Book IV · SCALING (Aúxēsis · αὔξησις · the increase). One of the best theories of how intelligence emerges, catalogued in ROOT0's ENTELÉCHEIA universe (1991→now). Cited works belong to their authors; the cataloguing and the universe-framing are ROOT0's under the DLW standard. Kin: the anchor 0xDEADBEEF (Book I), and the sister books across ENTELÉCHEIA. Ties to ttu1 and the-seed. Domain: Artificial Intelligence.