@hiljaisuus Ihmiset ovat siinä mielessä ihan tavallisia eläimiä että teemme aina sitä mikä on helppoa ja mukavaa. Kyllä #DanielKahneman’in #thinkingfastandslow -kirjan perusajatus pitää.

Consider engaging your careful thinking (System 2) when reading AI-generated text, as automatic thinking (System 1) can lead to quick trust. Being mindful of potential biases & limitations helps make informed decisions about AI output. #AI #ThinkingFastAndSlow #Suggestion

Several things you realize about billionaires/super wealthy after reading Thinking Fast & Slow - and are themes in #TheGlassOnion.
1) Many ultra rich are extraordinarily lucky, and their contributions would have likely come about due to natural progress.
2) Many sycophants and ultra rich themselves are often plagued by survivor bias and don't realize the insanely small odds of (1)
3) Because of (1) and (2) many of them are actually incredibly ignorant and attribute their success to natural talents. This often leads to them majorly fucking up when moving to any problem out of their domain. Not every musician is a historian. Not every business man is a politician. Not every engineer, is a software engineer
#billionaires #tech #Politics #thinkingfastandslow #statistics

This weekend will be exciting! John E. Laird (https://laird.engin.umich.edu/), Yoshua Bengio (https://yoshuabengio.org/) and @garymarcus are speaking at #AAAI #FSS 2022 on Thinking Fast and Slow - https://sites.google.com/view/aaai-fss22/program.

Laird and Bengio represent the cognitive-symbolic and statistical-sub-symbolic views respectively. Both have architectures that implement some form of 'fast and slow'. @garymarcus ofcourse is the chief cynic :)

#AI #ML #ThinkingFastAndSlow #CognitiveArchitecture #NeuroSymbolic

“样本越大，估算结果越准确”

“小样本比大样本更容易产生极端结果”

第一句话算是一个没有任何异议，任何人都会不假思索赞同的常识表述了吧。然而看到第二句时候，还是会忍不住停下来思考一会儿。其实道理也很容易想明白，一个箱子里放着同样数量得红球和白球，A每次随机取出4个球记录颜色，B每次随机取出7个球记录颜色，只要时间够长，A拿到同样颜色球的次数就一定比B拿到同样颜色球的次数多。用这个例子来说明就很容易理解第二个表述了吧。

一项研究对美国肾癌发病率进行调查，调查显示，人口稀少的、共和党地盘的乡村发病率最低。你会如何解释这个现象？很多人直觉就会是：乡村空气更好、污染更少、致癌物更少、生活压力较小，因此致癌率低。而现在，假设发病率坐高的县，也是人口稀少的、共和党地盘，大家又会如何解释呢？也很容易：乡村医疗条件差，三高饮食，养生意识差，大家嗜烟嗜酒，自然致癌率高。
那么问题来了，怎么人口稀少，既可以成为发病率低又可以成为发病率高的理由呢？

这时候我们就可以回顾下上文的样本大小的规律，用这个统计学原理解释这个现象。尽管我以为我已经融会贯通了“大样本比小样本准确”这个常识，但是在实际应用时，我还是很难用它联想、解释观察的现象，甚至在上文中的第二个表述我都得思考一会儿才明白。

我觉得这就是很多人学习时候遇到的问题，我们以为我们已经彻底了解了某个原理，但是直到你遇到某个问题之前，你永远不知道你究竟掌握了多少。如果你不知道你不知道，你要怎么知道呢。

#til
#读书笔记
#thinkingfastandslow