Introduction

In 1987, Nobel laureate Robert Solow observed that despite the explosion of computing technology, productivity growth had stalled—a phenomenon now known as the "Solow Paradox." Fast forward to 2026, and we are witnessing Solow’s Paradox 2.0. Companies are doubling their AI spend, mandating the use of generative tools, and aggressively cutting headcount under the guise of "AI-driven efficiency."

But the data is showing a different story. As explored in this insightful video, the rush to replace human workers with AI is resulting in a new, corrosive phenomenon: AI Work Slop.

The Cost of the "Efficiency" Obsession

The corporate narrative is uniform: Humans are expensive; AI is cheap. However, real-world implementations have consistently backfired.

  • Klarna’s Support Meltdown: Replacing 700 customer service agents with a chatbot initially boosted "efficiency" metrics, but customer satisfaction plummeted.
  • The Air Canada Fabrication: The legal fallout from an AI chatbot inventing a bereavement fare policy highlighted the dangers of corporate due diligence being replaced by blind faith in model output.
  • McDonald’s Tech Failure: Three years of testing AI drive-thru ordering only resulted in order errors and speed-of-service degradation.

The Rise of "Work Slop"

We are dealing with AI Work Slop inside our internal systems. Emails, project documentation, and slack messages are being flooded with AI-generated filler that "looks correct at a glance" but lacks substance. For the advanced engineer, this is a crisis of signal-to-noise. We are spending more time auditing AI hallucinations than we are building features.

Augmentation vs. Replacement

The video touches on a vital distinction for software engineers: Augmentation vs. Replacement. While GitHub data shows that nearly 27% of production code is now AI-authored, the entry-level hiring pipeline is collapsing. By automating away the "junior" tasks, we are inadvertently killing the training pipeline that produces the senior engineers of tomorrow.

Conclusion

As software engineers, we must move from being "passengers"—accepting AI output blindly—to "pilots"—using AI to extend our knowledge and skills while retaining rigorous human oversight. The future of engineering won't be won by those who produce the most "slop," but by those who can architect reliable systems that leverage AI for efficiency without surrendering their critical thinking skills.