A key Linux memory-management reviewer is stepping back, citing a flood of AI "slop" patches
Lorenzo Stoakes says low-quality AI-generated submissions have grown from an occasional nuisance to roughly half of what lands in his inbox.
The Linux kernel has lost some of its most valuable reviewing time to AI-generated patches. Lorenzo Stoakes, a lead reviewer for the kernel's memory-management code, said on the Linux kernel mailing list that he's stepping back from several of his review responsibilities, and he pointed to the volume of low-quality AI submissions as the underlying cause.
What he's giving up
Stoakes is leaving the core, reclaim, Rust, transparent huge pages (THP) and miscellaneous sections of memory management, while keeping his other duties. By Tom's Hardware's account the move is amicable: he plans to finish the patch series he's working on and says he'll still pitch in occasionally. The immediate reason he gave was the pressure on his inbox.
The AI problem
He was blunt about what's driving that pressure. He described the AI wave as getting significantly worse over time: at first he saw "slop" now and then, then weekly, then every few days, and now it makes up about half of what he sees, if not more. He had raised concerns about a wave of AI-generated patches in memory management as far back as July.
Every patch has to be read and judged by a human before it's accepted or rejected, so a flood of plausible-looking but poor submissions eats directly into the time of a small group of experts.
Why the kernel is a bellwether
The Linux kernel is one of the best places to watch AI's effect on programming, because its code is critical and its review standards are strict. AI tools can genuinely help: they make it easier for anyone to find and fix a bug, and the kernel's leadership has been broadly open to them. Under current practice, AI-assisted patches don't have to be labeled as such.
That openness shifts the burden to filtering. Open-source projects rarely have enough maintainers or funding to begin with, and when anyone with an AI agent can generate patches at scale, reviewer attention becomes the bottleneck. Stoakes's step back is the latest in a series of open-source departures, policy changes and new submission rules prompted by AI-generated contributions, and it's unlikely to be the last.
Sources
This story is based on reporting from these outlets. Read them for the full picture:
