Evan X. Merz

AI slop is the latest hot framework. It's the latest trend in software development that is eating an entire industry. Every executive in tech bought into the hype from people heavily invested in AI, and they have, at this point, fully swapped out a beautiful, hand-crafted codebase in favor of poorly written AI slop code bloat that is much more difficult to maintain.

Replacing carefully engineered systems with ai generated code that looks correct is a net negative for a business. It results in more downtime, harder to fix issues, code bloat, poor user experience, and poor performance. Furthermore, using AI coding agents trains your engineers to be lazy and rely on AI, which compounds the problem.

The AI slop hype train reminds me most of the trend in the 2010s toward noSQL databases such as MongoDB. The hype was so strong that smart, well-educated tech leaders were building companies from scratch based on noSQL databases. Of course a few years into the hype, the research showed that noSQL databases are outperformed at their own game by mature relational databases with json blobs. Many tech startups died because they realized too late that a traditional database would fit their use cases better than their massive noSQL cluster.

In short, tech companies jumped way too quickly on the noSQL hype train, and some of them paid the price. The same thing happened with nodejs, cryptocurrency, and even apps to some degree (is it really worth it for your company to have an app?). Companies that jumped on the hype train too quickly were punished for it. Companies that continued to rely on traditional engineering continued operating reliably.

The companies that have so quickly swapped out their human crafted codebase for AI slop are going to pay for it. It reflects bad leadership. If a company can't see that AI slop is not engineering, then the leadership doesn't understand the industry that they are leading.

I don't have the numbers to back this up yet. I only have the experience of watching this trend unfold at two of the biggest tech companies in the world. But the Microsoft outage reflects poor engineering culture that is too heavily reliant on AI. The quick enshittification of various Amazon products reflects their overly quick adoption of AI slop.

It might not be obvious now, but the numbers will soon show that traditional engineering is better for a company's bottom line than AI slop.