我始终相信自我迭代的力量。
我是一名 70 后，参加工作后攻读电子工程夜大，后来又自考法律本科。直到 11 年前老来得子，我才触及到人生最难的学科——教育。看着如今 11 岁的孩子身处「内卷」旋涡，我深切体会到，父母的认知才是孩子的天花板。如何在不确定的时代里，做好确定性的家庭教育，成了我当下倾注全力去求解的终极难题。
I have always believed in the power of iterating on oneself.
Born in the 1970s, I studied electronic engineering at evening college after starting work, and later took a law degree by self-study examination. It was only when my son was born, eleven years ago, that I met the hardest subject of all — education. Watching an eleven-year-old grow up inside the involution spiral, I have come to feel it plainly: a parent's own understanding is the ceiling above the child. How to do family education with some measure of certainty, in an age that offers none, is the problem I am now giving everything to.
上面那道题——如何在不确定的时代里做好确定性的家庭教育——就是这三篇文章的来处。它们横跨三个看起来毫不相干的场景——一间数学课堂里沉默的几秒钟、一个把手机攥得死紧的青少年、一套毫秒级给出对错的自适应学习系统——但它们追问的是同一件事：一道本应由一个人自己走完的发生工序，是怎么被别的东西替他走完的。
在课堂上，替他走完的是教师那套高速运转的完形程序：教师用自己成熟的认知结构去「看懂」学生尚未成型的混乱，在头脑中替他补好缺口，再基于这个已被补全的模型设计下一个「引导性」提问——被拿走的是学生独自做出未确证抉择并承担其重量的那个瞬间。在手机那一头，替他走完的是数字世界：当现实中一次完整的「发起—承担后果—确认这是我做的」都难以发生时，那个能够行动的自我第一次出生的地方，就不在现实里了——他不是逃去了虚拟世界，他是从来没有在别处出生过。在算法平台上，替他走完的是验证信号本身：从落笔到被判定对错的那段间隔一旦趋近于零，学习就从指向问题，转为指向验证。
三篇文章各自命名了一个此前没有名字的东西：代办与判断压力、生成寄生与活性湮灭、劳动-验证耦合间隔与认知及物性。它们的共同方法是：先撤销一条连批评者都还站在上面的先验，再让新的辨别面从矛盾里被逼出来，最后交出一组可以把自己判错的条件——三篇都在文末给出了能让作者认输的实验设计。
These first three essays span a silent few seconds in a mathematics classroom, an adolescent gripping a phone, and an adaptive learning system that returns a verdict in milliseconds. They pursue one question: how a step that a person should have walked himself gets walked for him — by a teacher's fast pattern-completion, by a digital world that becomes the birthplace of an acting self, or by a verification signal that arrives before the labour is finished. Each essay names something previously unnamed, and each ends with the experiment that would prove its author wrong.
观海听涛于 2026 年 8 月交来第一批稿件共十篇，经全文盲评后选发三篇，其余七篇未发。编辑部把择优的结果与理由一并公开，是本站的规矩：三篇之中，两篇为原稿加固后发表（增补部分不计入分数），一篇为四篇原稿的编辑合并稿（因四篇共用同一骨架，分开发表会造成互相碰撞）。未发的七篇及其理由，见作品页的编者附记。
Ten manuscripts were submitted in August 2026; three were selected for publication after full blind review, and the reasons for both the selection and the omissions are published alongside the work.