Shopify CEO Warns AI 'Slop Grenades' Are Creating More Work, Not Less
Key Takeaways
- •Shopify CEO Tobias Lütke warned that careless AI use can lower productivity, labeling unreviewed outputs such as flawed emails and code "slop grenades" that transfer extra work to colleagues.
- •Lütke's latest comments shift focus from encouraging AI adoption, which he earlier framed as a baseline expectation, to holding employees responsible for verifying AI-generated results.
- •A survey by BetterUp Labs and Stanford's Social Media Lab found that 52.7% of U.S. desk workers reported sending workslop, while 38% reported receiving it.
- •Recipients of workslop spent an average of 3.4 hours per month revising it, and the phenomenon reportedly damaged perceptions of senders' competence and reduced willingness to collaborate.
- •The data indicate that unreviewed AI output imposes hidden cleanup costs that could offset productivity gains for companies broadly pursuing AI adoption.

Shopify CEO Tobias Lütke is warning that careless use of artificial intelligence can reduce productivity rather than improve it, saying employees are circulating unexamined emails and code that colleagues must then review and repair.
Lütke calls such outputs "slop grenades" — work that appears finished on the surface but creates extra labor for whoever receives it. In an interview, he said AI should help employees synthesize information instead of producing long, unchecked messages that waste readers' time.
The criticism marks a shift in tone from Lütke's earlier push for employees to treat AI use as a baseline expectation, a stance he laid out in a widely shared internal memo. His latest comments focus less on whether workers use AI and more on whether they take responsibility for the results — a distinction that determines whether AI saves time or simply moves work to someone else's desk.
His warning ties into a broader debate over "workslop," a term researchers use to describe polished-looking AI output that lacks the detail needed to be actually useful. A survey by BetterUp Labs and Stanford's Social Media Lab found that 52.7% of U.S. desk workers reported sending workslop to colleagues, while 38% said they had received it and spent an average of 3.4 hours per month revising it. The figures suggest the phenomenon Lütke describes extends well beyond Shopify.
Typical examples include structured emails with broken links and unnecessarily complicated code. Recipients also reported that workslop damaged workplace relationships, making senders appear less competent or friendly and reducing their willingness to collaborate with them.
Lütke's remarks reflect a growing concern among technology leaders: AI can boost output quickly, but without review and accountability it can also shift hidden cleanup costs onto everyone else. For companies pushing AI adoption, the open question is how to keep the speed of AI-generated output from being eroded by the time others spend fixing it.