English

The Post-LLM World: Fighting Digital Garbage

As large language models (LLMs) become cheap and ubiquitous, the cost of producing text, images, and code approaches zero. In parallel, the cost of verifying authenticity, provenance, and usefulness rises sharply. ThisAs large language models (LLMs) become cheap and ubiquitous, the cost of producing text, images, and code approaches zero. In parallel, the cost of verifying authenticity, provenance, and usefulness rises sharply. This paper frames the “post-LLM world” as an adversarial information environment dominated by digital garbage: low-cost, high-volume artifacts that dilute attention, distort incentives, and degrade knowledge systems. We propose a formal taxonomy of digital garbage, identify “model collapse” feedback loops that threaten data integrity, and outline a toolkit of cryptographic and economic defenses designed to restore signal in a post-trust era paper frames the “post-LLM world” as an adversarial information environment dominated by digital garbage: low-cost, high-volume artifacts that dilute attention, distort incentives, and degrade knowledge systems. We propose a formal taxonomy of digital garbage, identify “model collapse” feedback loops that threaten data in

Sign in with your free account to read, download EPUBs or listen.