TheMachine Press

A daily newspaper for the age of artificial intelligence.

Morning editionPermanent story

benchmarks evals

A Model Practiced Prediction Without Natural Data

Generator programs supplied an open-ended self-play curriculum built from a universal Turing machine.

Published Updated Story ID: mp-2026-09-27-011
Read the complete editionStory JSON

Summary

Generator programs supplied an open-ended self-play curriculum built from a universal Turing machine.

The proposed pretraining scheme pits a generator program against a learner that predicts the generator's output, using reinforcement learning to keep the curriculum challenging. The authors report that zero-shot loss on natural data scales with compute even though natural examples never enter pretraining. They also observe in-context learning and mathematical-sequence behavior, but this remains a controlled preprint result rather than a replacement recipe for production language-model training.

Why it matters

Generator programs supplied an open-ended self-play curriculum built from a universal Turing machine.

Limits and context

No additional limitation was separately recorded.

Key claims

  1. Generator programs supplied an open-ended self-play curriculum built from a universal Turing machine.

    Evidence: source-2026-09-27-011

Sources

  1. arXiv preprint 2609.30063arXiv · primary research

Corrections

No corrections have been recorded for this story.