Thoughts on Slowing the Fuck Down
mariozechner.at
March 25, 2026
12 min read
🔥🔥🔥🔥🔥
76/100
Summary
Coding agents have advanced to the point where they can build complete projects autonomously, moving beyond earlier tools that served primarily as assistants. Many developers are now using their free time to create projects they previously lacked the opportunity to pursue.
Key Takeaways
- The use of AI coding agents has led to a decline in software quality, with many products exhibiting significant bugs and performance issues.
- Companies relying heavily on AI-generated code are experiencing increased brittleness in their software, with 98% uptime becoming common.
- There is a growing sentiment among developers that the rapid production of code through AI has resulted in a lack of discipline and poor design decisions.
- Many organizations are facing challenges due to the absence of code reviews and oversight, leading to features that are unnecessary or problematic.
Community Sentiment
Positives
- The core idea of evaluating the impact of AI tools on productivity is valuable, as it encourages a more thoughtful approach to software development.
- The advancements in AI models like Claude and Codex are impressive, challenging the notion that agent-written code cannot be production-ready.
Concerns
- Reliance on AI-generated code often leads to significant issues, such as memory leaks and crashes, highlighting the risks of overestimating AI capabilities.
- The sentiment that software quality has deteriorated despite technological advancements suggests a growing concern about the reliability of AI-generated solutions.
