Source: Theregister
The Linux kernel maintainer's observation suggests the initial wave of low-quality, automated code submissions generated by LLMs has crested—a relevant data point for open-source governance as projects figure out how to filter signal from noise. Rather than an endless avalanche of AI submissions, the pattern appears to be stabilizing into a more manageable baseline of bad patches, which means maintainers can now focus on sustainable vetting processes instead of panic-mode triage. The shift also tracks LLM adoption curves: the easy wins (blindly running AI on existing codebases for "contributions") burned out quickly once repositories tightened submission standards.