我一直描述不出来的东西是……

这句话来自我们的问卷。三十四个人接着它写下了自己的答案——雨来之前的气味、母亲叫我名字的方式、真心。现在轮到你。
写下它。一台机器会试着把它转译成一幅画面，风格和展厅里的十一段影像一样；它一定会误解一些什么——那正是这件作品想让你看见的。

This sentence comes from our survey; thirty-four people finished it before you. Now it is your turn. Write it down — a machine will try to translate it into an image in the style of the eleven films, and it will certainly misunderstand something. That is the point.

投稿先经过策展人的眼睛，再交给机器；生成的画面由策展人从多张备选中选定，标注为 generated second reading。不收集姓名；含他人可识别隐私的内容不会展示。
Submissions pass a curator before they reach the machine; images are chosen from several candidates and labelled as generated second readings. No names are collected; text that identifies other people is not shown.