打个比方，LLM像是“未出山前的诸葛亮”，善于分析，以“隆中对”和刘备对谈，出谋划策，但限于“纸上谈兵”；智能体则是“出山后的诸葛亮”，掌握全局情报，运筹帷幄，组织资源、调兵遣将，亲自率军北伐。
Not so much a bad night at the office as a high-stakes, avant-garde masterpiece of self-destruction, Ramy Bensebaini’s performance for Borussia Dortmund as they crashed out of Bigger Cup is destined to go down in the annals as one of the most hapless in the tournament’s history. While there have been costlier mistakes (hello, Loris Karius) and far more high-profile disintegrations (bonjour, b@nter-era PSG), it is difficult to recall any one elite professional footballer being responsible for quite so many howlers in one game as the hapless Algerian left-back.
，详情可参考safew官方下载
Harpreet Matharu，这一点在爱思助手下载最新版本中也有详细论述
Returning back to the Anthropic compiler attempt: one of the steps that the agent failed was the one that was more strongly related to the idea of memorization of what is in the pretraining set: the assembler. With extensive documentation, I can’t see any way Claude Code (and, even more, GPT5.3-codex, which is in my experience, for complex stuff, more capable) could fail at producing a working assembler, since it is quite a mechanical process. This is, I think, in contradiction with the idea that LLMs are memorizing the whole training set and uncompress what they have seen. LLMs can memorize certain over-represented documents and code, but while they can extract such verbatim parts of the code if prompted to do so, they don’t have a copy of everything they saw during the training set, nor they spontaneously emit copies of already seen code, in their normal operation. We mostly ask LLMs to create work that requires assembling different knowledge they possess, and the result is normally something that uses known techniques and patterns, but that is new code, not constituting a copy of some pre-existing code.，更多细节参见旺商聊官方下载