Unverified50% confidenceOpinionExact time
模型使用 bash 的成功率天然更高，因为它在预训练中见过大量真实命令行示例，而自定义工具 API 几乎不出现在预训练数据中
1
Sources
50%
Confidence
Long-term
Relevance
7/10/2026
First Seen
Sources
Files over Tools: Rethinking AI Agent Architecture with a Virtual Filesystem
hackernewshackernews7/9/2026
Related Claims
Unverified针对代码任务专项强化训练的模型在系统操作场景中表现显著优于通用对话模型72% similarUnverifiedAlignment training optimizes the model's statistical behavior in known scenarios rather than establishing formally verifiable security boundaries, and may fail when faced with rare instruction combinations or role-playing frameworks outside the training distribution.70% similarVerifiedThrough higher-quality training data, more optimized training strategies, and more refined post-training alignment, smaller-scale models can fully surpass larger-scale but lower-quality-trained models on practical tasks.69% similarUnverifiedOther residual connection improvement schemes outperform the original version in the early stages of training but fall behind the original version in the later stages.69% similarUnverified完整训练代码并不等于能立即产出媲美商业模型的成果，模型质量还取决于数据规模质量和算力持续投入69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/459589API
curl https://kongchang.com/api/v1/knowledge/claims/459589MCP
get_claim(id=459589)