AI 中文总结
本研究提出AI辅助开发中存在集成瓶颈现象，通过案例分析和实验验证，发现该瓶颈源于贡献无法转化为可执行指令，且现有机制无法满足所有交互要求，值得专门研究。
AI 中文摘要
AI辅助编程的研究一直聚焦于执行鸿沟，即用户如何编写成功的提示词。本文报告了一种候选现象——集成瓶颈，它位于Norman提出的评估鸿沟中：与修复相关的贡献到达用户后，在接收时无法转化为可执行的操作。在18个公开共享的AI辅助开发账户组成的语料库中，有2个案例报告了这种情况，分别来自同行和系统自身的输出；两者都发生在评估的解释阶段，而第3个案例本应发生在比较阶段，只有通过推断解读才会被发现，并被报告为边界案例。案例内的对比表明，可执行性取决于该贡献是否可以被重新表述为一条指令，而无需中间判断。本文将此作为一项值得专门研究的候选现象，而非既定规律；其证据基础是作者的回顾性自我报告。在提交方面，仅媒介无法区分语料库的对比；在4个决定性对比中，即时采纳与可核查的条件一致，而第5个案例显示，足以实现即时采纳的可核查性并不能保证持续性。针对两个当前模型进行的20次多轮探测，在15个可判断的窄探测中未观察到约束损失，且在8次带有明确重构请求的10次扩展探测中也未观察到约束损失。研究2并非复制尝试：它分离了语料库作者所援引的机制——再生，并证明该机制不足以解释相关现象。本文为公认约束清单指定了4个交互要求；对12种机制的一致性分析发现，没有任何一种机制被记录为满足所有4个要求。
英文摘要
Research on AI-assisted programming has concentrated on the gulf of execution -- how users write successful prompts. We report a candidate phenomenon, an integration bottleneck, that lies in Norman's gulf of evaluation: a repair-relevant contribution reaches the user and fails to become actionable at the point of receipt. Two cases in an eighteen-case corpus of publicly shared AI-assisted-development accounts report this, from a peer and from the system's own output; both fall at evaluation's interpretation stage, and a third, which would fall at comparison, is reached only on an inferential reading and reported as a boundary case. A within-case contrast is consistent with actionability turning on whether the contribution can be restated as an instruction without an intervening judgement. We report this as a candidate warranting dedicated study, not an established regularity; its evidence base is retrospective author self-reports. On the submission side, medium alone did not sort the corpus contrasts; immediate uptake aligned with a checkable condition across four decisive contrasts, while a fifth case shows checkability sufficient for immediate uptake did not guarantee persistence. A twenty-trial multi-turn probe across two current models observed no constraint loss in fifteen judgeable narrow trials, and a ten-trial extension with an explicit restructuring request none in eight. Study 2 is not a replication attempt: it isolates regeneration, the mechanism the corpus authors invoke, and removes it as a sufficient explanation. We specify four interaction requirements for an accepted-constraint ledger; a conformance analysis of twelve mechanisms found none documented to satisfy all four.
CommentsThis paper has been withdrawn by the author. The analysis reported in Sections 4-5 was based on a preliminary version of the coding scheme, and the counts and reliability figures presented there are not sufficiently reliable to support the paper's claims. A corrected study is in preparation
