Status Updates From Myśl krytycznie i nie daj s...

驗證性偏誤是一種察覺、相信，並分享符合你我既存信念之資訊的傾向。當某種說法肯定我們對世界的信念，我們就會更傾向接受該說法為真，比較不會質疑其可能為假。
社會學家尼爾 • 波茲曼（NeilPostman）的格言一語中的：「在任何時候，你得應付的主要鬼扯消息來源就是你自己。」

Chpt 3 the nature of bullshit
To persuade or impress not to lead to truth
Latour’s blackbox many bluffs and cites. To call one bullshit call all of them bullshit

criminal machine learning:
E.g 2016, ArXiv detect criminal from facial features. Data problem, non criminal smiling professional headshots and criminal Id photos. Curvature of the mouth etc just cause of data

Chapter 2 Media Msg and Misinformation
Incentive misalignment of fostering long term subscription versus getting clicks
“Make u feel bla” most popular headlines
Share to signal self versus truth
Repercussions e.g. 2018 fake videos abt kidnapping, xenophobia, killed tourists
E.g Jordan peele and Obama face swap. Macedonian children and Russian bots. When ppl are scared enough, they stop talking.

Chpt 1 Bullshit everywhere
Brandolini’s principal of creating bullshit is so much easier than cleaning it up. E.g Wakefield’s anti vaccine paper and documentary

And bullshit spreads faster

相關性不代表因果關係

有效的鬼扯，就像是說話者弄了一個黑盒子。如果是純粹謊言而已，通常很容易便能加以揭穿及反駁，但若鬼扯得好，便不易查證是否屬實。就像拉圖所說的黑盒子，鬼扯能夠替某種說法找掩護，不受進一步調查的威脅。