Shopify CEO Tobias Lütke Warns AI 'Slop Grenades' Are Sabotaging Productivity in 2026

What Are 'Slop Grenades'? Shopify's CEO Coins a New Problem for the AI Workplace

"Slop grenades" is the term Shopify CEO Tobias Lütke recently coined to describe AI-generated content or code that employees produce quickly and without proper review, effectively transferring the burden of verification to the recipient. The phrase emerged from Lütke’s own experience leading a company that has aggressively embraced artificial intelligence, and it has quickly resonated across the tech industry as a cautionary tale about the hidden costs of AI adoption.

In an interview and subsequent company-wide discussion, Lütke illustrated the concept with a simple example: an employee uses a large language model to draft a lengthy email, then sends it without editing or condensing. The recipient then has to paste that email into another AI tool just to summarize it, effectively creating a cycle of unnecessary work. "You're not saving time; you're just moving the burden," Lütke told Fortune. The same dynamic applies to code: an AI-generated pull request that hasn't been reviewed or tested simply forces a colleague to debug and validate, turning a productivity gain into a productivity tax.

The term has generated immediate buzz because it comes from a CEO who has been one of the most vocal proponents of AI in the workplace. Lütke has publicly pushed Shopify to become an "AI-first" company, mandating that employees prioritize AI tools in their daily workflows. Yet even he now acknowledges that the technology introduces a new class of friction that leaders must address.

How Shopify's Own AI Mandate Created the Slop Grenade Problem

To understand why slop grenades emerged at Shopify, it helps to look at the company's AI strategy. In early 2025, Tobias Lütke issued a company-wide mandate requiring employees to prioritize the use of AI in their work, from writing code to drafting internal communications. The centerpiece of this push is River, a custom internal AI agent built on large language models that Shopify has trained on its own codebase and documentation. According to StartupFortune, "approximately half of Shopify's production pull requests now originate from conversations with River." That is a staggering statistic, reflecting how deeply AI has been woven into the company's engineering culture.

But the speed of AI-generated output has outpaced the company's internal review culture. When employees can generate code or text in seconds, the temptation to ship without thorough human oversight becomes strong. Lütke himself admitted in a podcast interview featured on The Fabulous that the true cost of AI productivity often lies in the review burden shifted to others. "Human judgment and responsibility remain crucial," he emphasized. "The AI is a tool, but the human is still accountable."

The problem isn't that the AI-generated work is always bad — much of it is quite good. Rather, the issue is that unreviewed output, even when mostly correct, introduces risk and cognitive load for the person who has to verify it. An AI-generated email might be mostly accurate but contain a subtle factual error; an AI-generated code commit might compile but introduce a security gap. The recipient cannot simply trust the output; they must scrutinize it, often spending as much time verifying as they would have spent writing the original from scratch.

Perspective	Core Claim	Implied Solution
Tobias Lütke (Shopify CEO)	Slop grenades shift work to coworkers; employees must review AI output before sharing.	Humans must remain accountable; build review checkpoints into AI workflows.
Mitch Joel (Commentator)	Slop is a human accountability problem that predates AI; AI just makes it easier to create.	Fix the culture of passing off unreviewed work, regardless of the tool used.
Industry Observers	The productivity gains of AI are real, but the verification tax is often hidden until it reaches a tipping point.	Companies need clear policies on when AI output requires human sign-off.

The Human Factor: Mitch Joel Argues Slop Grenades Aren't AI's Fault

Not everyone agrees that AI is the root cause of slop grenades. In a widely circulated Six Pixels column, commentator Mitch Joel pushed back on the framing. "AI didn’t create work slop… we did," he wrote, arguing that the tendency to pass off unreviewed, half-baked work to colleagues has existed long before generative AI. "We've been sending poorly thought-out emails, sloppy code, and half-finished memos for decades. AI just makes it faster to create that slop, so the volume has increased."

Joel’s argument reframes the problem not as a technical bug to be patched but as a cultural failure of accountability. If an employee would never send a colleague a rough draft without first cleaning it up manually, why would they do so when the draft was generated by AI? The answer, Joel suggests, is that AI lowers the perceived cost of creation, making it feel less like "work" and therefore less in need of review. He warns that companies rushing to adopt AI without addressing this cultural gap will simply see their slop grenade problem scale with every new AI tool they deploy.

This perspective adds nuance to Lütke's own warnings. While the Shopify CEO focuses on the practical burden placed on coworkers, Joel highlights the deeper behavioral issue: the tool is not the culprit; the choice to skip review is. Both agree, however, that the outcome is the same — more work for others and a net drag on productivity.

Practical Implications for Companies Using AI Agents

• Establish mandatory review gates for AI-generated work. Before any AI draft — whether code, email, or report — reaches a colleague, the creator must review and edit it. Lütke himself has called for this in internal guidance at Shopify.
• Train employees on verification techniques. It's not enough to say "review the output"; workers need to know how to efficiently audit AI-generated content for errors, hallucinations, and omissions. Shopify has started offering internal workshops on prompt engineering combined with quality assurance.
• Measure both creation time and review time. Most productivity metrics today capture how fast something is produced, not how long it takes others to consume or verify it. Search Engine Journal reports that experts are now calling for companies to track "net productivity" — the difference between time saved by the creator and time spent by the reviewer.
• Set a cultural norm of accountability. Leaders must model the behavior of reviewing and refining AI output before sharing. If executives send slop grenades, junior employees will follow.

The Future of AI at Work: Balancing Speed and Review

Shopify’s experience with slop grenades is not unique — it is a harbinger of challenges that many organizations will face as AI agents become standard tools in the workplace. The very features that make AI powerful — speed, ease of generation, and language fluency — also make it easy to produce work that looks polished but hasn't been sanity-checked.

The key insight from Lütke’s candid admission is that productivity gains from AI are not automatic. They require deliberate orchestration of workflows, clear ownership of quality, and a culture that values careful review over fast output. Shopify is now building these guardrails into its processes, treating River not as a replacement for human judgment but as a first draft generator that always needs a human editor.

Some companies may respond by restricting AI use or requiring approval for AI-generated code, but Lütke has not called for a retreat. Instead, he continues to advocate for AI adoption — with the caveat that every team must confront the slop grenade problem head-on. As he told Business Insider, "The solution isn't to stop using AI; it's to stop treating AI output as final."

The lesson for the broader tech industry is clear: AI can dramatically accelerate creation, but the bottleneck of verification must be accounted for in any productivity equation. Companies that ignore this risk drowning in their own slop grenades.