The future of everything is lies, I guess: Work
aphyr.com
April 14, 2026
12 min read
🔥🔥🔥🔥🔥
59/100
Summary
Software development may increasingly resemble witchcraft rather than traditional engineering. The rise of AI coworkers raises concerns about the robustness of systems, as automation can complicate new domains.
Key Takeaways
- Automation and machine learning may lead to deskilling, automation bias, and monitoring fatigue in software development.
- The use of large language models (LLMs) for programming is increasing, with some companies relying on LLMs to generate all their code.
- LLMs may not preserve the reasoning properties expected from traditional compilers, leading to potential issues with software correctness.
- The future of programming may involve a shift towards using natural language for code generation, likening programmers to "witches" who invoke LLMs to create software.
Community Sentiment
Positives
- The ongoing discussion about AI's potential suggests that we may still be at the beginning of significant advancements, indicating a promising future for model capabilities.
- The idea of stacked sigmoid curves implies that even if individual advancements are modest, their cumulative effect could lead to substantial improvements in AI applications.
Concerns
- Concerns about the limitations of LLMs suggest that we may be nearing the peak of their technological improvement, which could stifle future innovation.
- The reference to risks in AI, such as those highlighted by the Air France 447 incident, underscores a critical need for better regulation and safety measures in AI deployment.
