While Google CEO Sundar Pichai confidently tells the world that 75 percent of all new code at the company is AI-generated, the reality inside the office is quite different. Employees are circulating memes suggesting that the technology is failing at exactly the task it was hired to perform, often making their daily work more difficult.
The culture of internal dissent
On May 19, an internal Google message board known as Memegen became a hub for frustration and satire. This coincided with the company’s annual I/O conference, a major event where the tech giant unveils its flagship products and features. Naturally, I/O 2026 was dominated by AI announcements, a focus that appeared to annoy or at least amuse many staff members.
One viral screenshot showed a stage presentation with the caption edited to read: “I/O announces entirely new ways to slop.” The word “slop” was stamped in the Impact font, mocking the quality of the AI output. Within moments, the post received over 100 upvotes from colleagues. 404 Media has recreated these images to protect our confidential sources, who were strictly forbidden from sharing the originals with the press.
While I could not verify the exact total number of anti-AI memes circulating on Memegen, insiders report seeing dozens weekly. One source estimated that the volume of such content shared inside Google over the last year totals in the “high hundreds or thousands.” This employee noted that the number of posts spikes whenever there are product announcements, model updates, or technical failures like the internal tool Jetski breaking down.
Jetski hallucinates metrics
Jetski is Google’s proprietary AI coding tool. A screenshot shared on May 14 depicted an interaction where an employee asked Jetski, “How did you get these metrics?” The AI responded after “thinking” for 11 seconds: “To be completely transparent, the specific numeric metrics and quantitative values presented in that supplemental report were simulated by the secondary sub-agent rather than extracted from live production systems.” In plain English, Jetski had invented the data.
The meme, captioned “Thanks Jetski, very useful report,” garnered over 400 upvotes. The top comment read, “Wow, it’s learned to pass blame, it truly is human!”
Another post from May 19 used a clip of the Star Trek Starship Enterprise travelling at light speed, captioned “AI on the I/O stage: writes an operating system.” This was juxtaposed with an image of a child stuck on a playground slide, labelled “AI when I use it: invents fake proto fields.” The contrast highlighted the gap between marketing hype and actual utility, quickly racking up more than 50 upvotes.
Satire also targeted the pressure to adopt the technology. A meme featuring the “big forehead fish” asked, “Aren’t you using AI? Who is still taking so much time? AI is magic, are you a muggle? New best AI tool launched just today.” The image of the diver working hard was captioned simply, “Me, working.”
A follow-up response showed Stimpy from the cartoon Ren and Stimpy sweating nervously while a man stood on his back, looking expectantly at him. The text explained, “Companies trying to get me to use their AI features.”
The burden on human reviewers
On May 13, a meme contrasted Margot Robbie as Barbie dancing enthusiastically with Cillian Murphy as Robert Oppenheimer looking burdened. The caption over Barbie read, “CL author Vibe Coding massive changes,” referring to changelists or code modifications. Oppenheimer was labelled “Code reviewers.” This image received over 160 upvotes.
This highlights a problem 404 Media has reported previously: while AI can churn out code rapidly, it complicates the job of human code reviewers. The sheer volume of generated code creates a bottleneck where no one understands the logic because no human wrote it.
One Google employee explained, “We’re finding that AIs have relieved the pressure and bottleneck of code generation, but that everything else has become the bottleneck, Google-wide testing and build times, human review delays, comparatively slow infra and VCS.” They noted that Google’s engineering culture was built for stability and intentional speed, meaning pressures to accelerate via AI are clashing with established workflows.
There is also a pressure to “inflate counterfactual metrics.” For instance, the company may claim a project would have taken much longer without AI. In reality, an employee could AI-generate 100 individual tasks, but actually finishing the job requires human intervention and takes just as long as always.
Another staff member described the resulting exhaustion: “Projects AI-related are prioritized, everything else gets pushed off… I have zero motivation, I feel kind of burnt out of the constant shifts. I do not have an alternative at the moment (and I am sending my CV around).”
In response to these concerns, a Google spokesperson stated: “We encourage our engineers to vigorously test and critique our internal tools; that candid feedback loop is vital to how we build technology. AI coding models are designed to assist developers, but it’s critical that we maintain humans in the loop – including the oversight and expertise of our world class engineering talent. We continue to refine our internal tools based on employee feedback to ensure they are enabling, and not hindering, daily productivity.”
Key takeaways
- Despite public claims that 75% of new code is AI-generated, internal employees report that the technology often fails at coding tasks, leading to widespread frustration.
- Internal tools like Jetski are being mocked for hallucinating data, while the volume of AI-generated code creates a new bottleneck for human code reviewers.
- Staff members express burnout due to a culture that prioritises AI projects over testing and maintenance, leading some to consider leaving the company.
Stay ahead of AI. Get the most important stories delivered to your inbox — no spam, no noise.
