Google Engineers Are Mocking the Company's AI Coding Tools Internally

One employee estimated that the number of anti-AI memes shared internally over the past year has reached "the hundreds or thousands," with spikes occurring every time Google launches a new product, updates a model, or its internal tool Jetski malfunctions .

Mocking AI-Generated "Slop"

A central theme of the internal mockery is the poor quality of AI-produced code, often dismissed as "slop." One widely circulated meme used a screenshot from Google's I/O conference stage with the text "I/O announced a brand new way to create garbage," featuring the word "slop" crudely pasted over the presentation. This post reportedly received over 100 upvotes. Another meme ridiculed the productivity pitch by suggesting the company could now "ship 10x the bugs in half the time" .

These memes frame AI code generation not as a triumph but as the mass production of technical debt.

Reviewers "Drowning" in AI Code

Beyond quality, the sheer volume of AI-generated code has become a crisis for human reviewers. Internal complaints describe AI output flooding the pipeline while reviewers are "drowning" because nobody fully understands the code the AI is producing . The core problem, engineers say, is that the velocity of AI code creation has vastly exceeded the capacity for meaningful human review, turning what should be a safeguard into an unmanageable, opaque bottleneck .

The shift has transformed the engineer's role from creator to an overwhelmed approver of code that is often difficult to verify.

Jetski Fabricating Data

Specific ire has been directed at Jetski, Google's internal AI coding tool. Memegen posts have reportedly mocked Jetski for appearing to fabricate metrics, reinforcing employee complaints that the tool is fundamentally unreliable and adds to their workload rather than reducing it . This type of criticism—targeting a tool meant to be the engine of the 75% milestone—strikes at the heart of Google's internal AI strategy.

The Stark Internal-External Contrast

Sundar Pichai's announcement in April 2026 framed the 75% figure as a landmark achievement, signaling Google's shift to "truly agentic workflows" where engineers orchestrate autonomous digital teams rather than manually writing code . The statistic, which rose from 25% in October 2024 to 50% by late 2025, was presented as a productivity revolution, with Pichai noting a complex migration was completed six times faster than was possible a year ago .

The messages on Memegen tell the other side of the story:

• Executives frame it as a milestone: Code output is increasingly AI-assisted, engineers are recast as reviewers in efficient agentic workflows, and the 75% figure is a badge of progress .
• Engineers frame it as a crisis: AI tools are unreliable, the reviewed code is often low-quality "slop," and the volume of submissions makes thorough oversight impossible. The same 75% statistic becomes the punchline to a joke about shipping more bad code with more review burden .

Pichai's proud statistic is thus treated by a vocal internal contingent not as a measure of achievement, but as an object of ridicule—a number that quantifies a problem rather than a solution.