Why do models trained on today's web get progressively worse as AI writes more of it?
機会
The web is now the primary training corpus for frontier models and is already saturated with AI-generated text that no deployed filter reliably catches. Research published across 2024 to 2026 shows that even a fraction of a percent of synthetic data in a training run triggers distributional collapse over successive generations, narrowing output diversity and degrading tail performance. The feedback loop is structural: models trained this year produce content that contaminates the corpus for next year's training run. Proposed mitigations such as source-level allowlists, watermark filters, and synthetic-data verifiers each have bypass vectors and none has been deployed at web-crawler scale. There is no agreed protocol for identifying and quarantining AI-generated training data before it enters a model.
重要な理由
A degrading shared training corpus sets a ceiling on every model built from public data, and that ceiling gets lower with each generation.
機会をどう評価するか
Opportunity Scoreは測定値ではなく、私自身の見解です。どれほど痛みを伴うか、どれほど頻繁に影響を与えるか、そして今日時点で解決策がいかに少ないか。スコアが高いほど、構築する価値が高いと私は考えています。
それが現れたときにどれほどの痛みをもたらすか。
実際にどれほど頻繁に人々がそれに直面するか。
今日時点で、それに対する優れたツールがいかに少ないか。