← Thursday, October 8, 2026Models & trainingBenchmark

A reproducible test explores whether LLMs can relearn telegraphese

fiveminutesforward.com 10 min read Editor score 6/10

The project provides a harness, frozen ledgers and a one-click notebook to study concise communication under telegraph-era per-word costs, comparing codebooks with another compression strategy. Reproducible tests of model behavior can help engineers assess how systems adapt to constraints.

What

A reproducible test compares codebook substitution and cablese instructions as ways to reduce LLM token use.

How

Across 50 passages and about 1,300 questions, cablese saved 48.4% of tokens with GLM-5.3-Flash and had 1.09 in-family recovery.

Why it matters

For AI system builders, the results show that token savings should be evaluated alongside answer recovery; the technique did not pay for every model.

Read on fiveminutesforward.com → Hacker News · 56 comments

Headline, summary, brief and diagram are written by a model from the source's own text; the link, byline and numbers come from the source. Read the original before relying on a claim.