Overview
- Tobias Lütke said Thursday on a podcast that careless AI outputs he called “slop grenades” are being passed between employees and are creating extra work for the people who must fix them.
- Lütke disclosed that Shopify’s internal agent, River, handles a large share of the company’s production code pull requests and that scaling such automation increases the risk of unchecked, low-value output.
- A survey by BetterUp Labs and Stanford’s Social Media Lab of 962 U.S. desk workers found 52.7% reported sending polished but flawed AI work and 38% reported receiving it, with recipients spending an average 3.4 hours per month revising it.
- Respondents said this type of output—called 'workslop'—looks legitimate but often omits key details or contains broken links and overly complex or incorrect code, and it makes senders seem less competent and less friendly.
- Executives who once pushed AI-as-baseline are now stressing human review, accountability, and governance to preserve real productivity and workplace relationships as companies expand internal automation.