No AI Slop: A Manifesto

Created by speakers and attendees of the 2026 AIE Code Summit, the No AI Slop manifesto argues that AI can greatly expand software teams’ capacity while making it dangerously cheap to produce work that people still have to review. It defines slop not as AI-generated material itself, but as work passed on without enough care for its quality or for the effort required of its audience. The manifesto warns that generated code, documents, and messages can overwhelm human attention; that poor material accumulates as later agents copy it; and that pressure to increase output can erode communication and quality. Its central values are outcomes over output, understanding over speed, human thinking over a model’s first draft, better work over easier work, and systemic fixes over repeated patches.

The proposed response is to keep people accountable for intent, judgment, and consequences, regardless of how much AI contributes. Teams should understand the problem before using models, disclose AI-written material when sending it, and check important parts according to the stakes. They should give colleagues kind, direct feedback when work creates unnecessary review burdens. Automated checks, agent reviews, linters, and type checkers can help, but do not eliminate the need for human review; when mistakes recur, teams should improve tools and guardrails. The manifesto also distinguishes harmful slop from useful shortcuts such as one-shot prompts or vibe coding: the test is whether the approach serves a useful product, maintainable system, and people’s time. It presents these principles as evolving guidance, not a settled formula.