问题
Why do we stay inside echo chambers?
把你认为相似的卡片归为一类。分类只为方便阅读，不决定任何事。
已关闭 · 返回行动阶段
正以访客身份查看。输入用户名即可加入并参与。
先按相似性形成类别，最后再命名
标准顺序：把你认为相似的想法两两关联——关联是对称的（A~B 即 B~A；方向无关紧要），一条关联在 2 名成员做出后成立。
为什么
关联是对称的——A~B 即 B~A。类别是从成立的关联中自然形成的分组，只有已形成的类别才能被人命名。你自己的关联带有 ✕，可以撤回；你不能移除别人的。类别只用于整理墙，从不限制投票——两张提出相同主张的卡片属于澄清阶段的合并。
关联和命名只能在归类阶段进行；主持人的 ← 可以重新开启它。
未命名类别 ——如果这一类有共同主线就为它命名，否则保持未命名 (2)
An A/B test measures engagement because that is what we can measure in a week
We cannot publish what we remove without teaching people to evade it
~ The people affected w…The people affected worst are in markets we do not staff (16)
~ An A/B test measures …An A/B test measures engagement because that is what we can measure in a week (18)
未命名类别 ——如果这一类有共同主线就为它命名，否则保持未命名 (2)
The rules are written in one country and applied in forty languages
~ The rules are forty p…The rules are forty pages and the decision takes twelve seconds (17)
The rules are forty pages and the decision takes twelve seconds
~ The rules are written…The rules are written in one country and applied in forty languages (17)
未命名类别 ——如果这一类有共同主线就为它命名，否则保持未命名 (2)
A moderator sees the worst of it for the wage of a call centre
~ Moderation is a cost …Moderation is a cost centre and ranking is a profit centre (16)
Moderation is a cost centre and ranking is a profit centre
~ A moderator sees the …A moderator sees the worst of it for the wage of a call centre (16)
· 尚未归类 (26)
The metric was chosen because it was easy to log
Nobody can explain why a particular post was shown
~ Nobody outside can re…Nobody outside can reproduce a single claim we make (13)
~ Regulation asks for t…Regulation asks for transparency reports nobody can verify (18)
Nobody is promoted for the feature they did not ship
An appeal is judged by the system that made the decision
A change that reduces engagement is reverted before anyone asks why
The founders were twenty-five and nobody senior said stop
Every external audit proposal dies in procurement
We built for a scale nobody had governed before
~ Nobody outside can re…Nobody outside can reproduce a single claim we make (12)
Every fix ships to the users who complain loudest
Ad money rewards volume of attention, not quality of it
There is no reviewer who speaks the language of half the reports
The people affected worst are in markets we do not staff
~ We cannot publish wha…We cannot publish what we remove without teaching people to evade it (16)
Safety work has no metric, so it has no headcount
Regulation asks for transparency reports nobody can verify
~ Nobody outside can re…Nobody outside can reproduce a single claim we make (18)
~ Nobody can explain wh…Nobody can explain why a particular post was shown (18)
The team that would study long-term effects does not exist
Harm shows up in a quarter and the review cycle is six weeks
Researchers ask for data and legal says no
Recommendation systems are not inspectable from outside
The model is retrained weekly and nobody diffs its behaviour
Ranking optimises for time spent, not for being informed
The people who raise concerns are moved to another team
The company would rather be regulated than be sued
People choose the sources that already agree with them
Automated removal is cheaper than being right
Nobody outside can reproduce a single claim we make
~ We built for a scale …We built for a scale nobody had governed before (12)
~ An A/B test measures …An A/B test measures engagement because that is what we can measure in a week (13)
~ Regulation asks for t…Regulation asks for transparency reports nobody can verify (18)
~ Nobody can explain wh…Nobody can explain why a particular post was shown (13)
Nobody agrees on who should decide what is true