AI Tribalism

3 months ago

The author initially dismissed LLMs as unsuitable for serious software development but now uses Claude Code for 90% of their coding tasks.
Debates around LLMs have become highly politicized and tribal, with opinions ranging from seeing them as revolutionary to destructive.
Despite initial skepticism due to early LLM limitations like hallucinations and bugs, significant improvements in 2025 made them indispensable for the author.
LLMs now assist in various tasks, including finding security vulnerabilities, writing benchmarks, and improving accessibility, though not perfectly.
The author acknowledges the fear and uncertainty among developers about the future of their profession but advises experimentation and open-mindedness.
Comments reflect a mix of resistance, cautious optimism, and enthusiasm about integrating LLMs into programming workflows.
The rapid evolution of LLMs suggests that adaptability and continuous learning are crucial for developers moving forward.

Hasty Briefsbeta