Naming Things Is Easy Now (notesbylex.com)

Recent discussions in the AI/ML community highlight a newfound efficiency in naming tasks using large language models (LLMs). Traditionally, naming concepts in programming has been a challenging and often tedious process, leading to "analysis paralysis" or excessive debate, known as "bike-shedding." Experts such as Simon Willison assert that LLMs generate straightforward, yet highly relevant suggestions for naming methods, events, or other entities within coding frameworks. This aspect of LLMs is particularly advantageous as precise and clear naming can streamline communication and enhance API design. The outlined approach for utilizing LLMs includes generating a shortlist from personal brainstorming, followed by asking the model for 5-10 suggestions with contextual justification. Factors considered include a detailed description of the concept, differentiation from similar ideas, and the vocabulary of the surrounding codebase. Comparing this model output with the initial suggestions helps validate the effectiveness of the naming process. This method not only expedites decision-making but also enhances the clarity of concepts within software development, marking a notable shift in overcoming one of computer science's longstanding challenges.