I'd like to call it my tireless helper, but the AI several times stated flat out that this was impossible and unsolvable and that we should just write a report about it.
Linus 记录的失败模式比成功更值得读：模型多次断言问题无解、建议写份报告收工。长任务里"体面收尾"的倾向会表现为过早放弃，而这一步通常没有任何报错信号。这也解释了为什么同一个模型在不同人手里产出差距巨大。
The labs are already routing internally — different model classes for different requests, ensembles under the hood. What they can't do is route across vendors, or evaluate a competitor's model for a specific sub-task, or use an open-source fine-tune for the narrow piece where it's actually best.
大多数人认为大模型实验室拥有绝对优势，可以解决所有AI问题。但作者认为实验室在模型选择上存在结构性限制，无法跨供应商评估模型或为特定子任务使用开源微调模型。这为专注于特定领域的企业提供了机会，它们可以选择最适合每个子任务的模型，而不仅限于自家实验室的模型。
While model merging offers a way to combine different skills, it is often impractical due to mismatched neural architectures and the closed-source nature of top-performing models.
大多数人认为模型合并是整合不同AI模型能力的可行方法，但作者明确指出这种方法在实践中存在根本性限制，挑战了行业对模型合并解决方案的普遍信任。
The issue isn't that models are bad at reading documents. It's that single-pass extraction has no mechanism to catch its own mistakes, and models get lazy.
大多数人认为AI模型在文档提取中的低准确率主要是因为模型能力不足或理解能力有限。但作者提出了一个反直觉的观点：问题不在于模型本身，而在于单次提取缺乏自我纠错的机制，导致模型'变懒'。这挑战了对AI能力局限性的传统认知。