Unverified50% confidenceFactExact time
Function Calling缺乏统一标准，不同模型提供商的接口和格式存在差异，切换模型供应商时代码往往需要大幅调整
1
Sources
50%
Confidence
Long-term
Relevance
7/22/2026
First Seen
Sources
Related Claims
UnverifiedThe semantic structure of code differs fundamentally from natural language, making it difficult for embedding models to accurately preserve information such as identifier naming, call relationships, and scope.70% similarUnverifiedLarge language models may confuse APIs from different versions when generating code, or omit import statements required for actual compilation.68% similarUnverified大语言模型每次对话独立，无法对整个代码库形成持久认知模型，导致不同会话生成的代码在架构风格、命名规范、错误处理策略上存在系统性不一致67% similarUnverifiedIn multi-Agent collaborative development, large language models have a tendency to be "overly enthusiastic" and may overstep their responsibilities by modifying code outside their designated scope.64% similarUnverifiedLarge language models have inherent non-determinism — even with identical prompt inputs, the model may return results that differ in format, content, or even logic at different times.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/593965API
curl https://kongchang.com/api/v1/knowledge/claims/593965MCP
get_claim(id=593965)