◉ PSYCHOHISTORY

Scaling Illusion

mechanism
AI's benchmark scores keep climbing while the actual output barely improves — a trick of measuring the wrong thing.
Who they are

The 'Scaling Illusion,' a mechanism the engine describes.

What they do

It's the idea that AI progress looks better on paper than it really is.

How it works

As more computing power is thrown at models, the internal metrics (loss curves) keep improving even while the real output quality stagnates — because people evaluate the metrics instead of the actual results.

Why it matters

It warns that impressive-looking numbers can mask diminishing real-world returns on all that energy and compute.

The engine's record — word for word
Diminishing returns on compute energy masked by metrics. Loss curves improve while output quality stagnates. Evaluating metrics instead of output.
Follow the trail
Walk this on the live map →
Part of the Psychohistory engine — 2,426 entities, 6,314 documented connections. Open data, built to be proven wrong.