Codex 的能力 Overhang
Whenever I don't use codex for a task, I ask myself why and usually realize that there's some missing context, I needed to write a skill, or I just di...
Whenever I don't use codex for a task, I ask myself why and usually realize that there's some missing context, I needed to write a skill, or I just didn't think to use it.
Rarely is it because the task is outside of the capabilities of the model. Overhang right now feels large.
Greg Brockman（OpenAI President & Co-Founder）分享了他使用 Codex 的反思：每次他没有用 Codex 完成某个任务时，问自己原因，通常会发现是以下三种情况之一：
- 缺少上下文 — 没有给 Codex 足够的背景信息
- 需要写一个 Skill — 没有把可复用的流程封装成技能
- 没想到用它 — 纯粹是习惯问题，没意识到 Codex 可以胜任
真正因为模型能力不足而无法完成的情况极少。 他认为当前的"能力过剩"（Overhang）感非常明显——模型能做到的远比人们实际使用的多。