Hell Fast
为了游戏化,LC和CF停滞了一年半 现在要回归啦, LC的目标是Guardian 2230 CF的目标是Stable Expert 1600


Hell Fast
Bitcoin is basically digital gold with highly consensus.

Hell Fast
现在 ai 的问题,不是解决不了人类解决起来难的问题 而是解决了太多人类都解决不好的问题


Hell Fast
Do not pray for easy lives. Pray to be stronger men. Do not pray for tasks equal to your powers. Pray for powers equal to your tasks. Then the doing of your work shall be no miracle, but you shall be the miracle. ― Phillips Brooks

Hell Fast
人当遇到复杂问题 会诉诸情绪和不理智 是对世界复杂度无奈和没有安全感一种妥协或是伪装

Hell Fast
共识形成慢,既为创业提高难度,也为垄断提供壁垒


Hell Fast
Working extremely hard works extremely well.

Hell Fast
Now I have all the energy to do the hard things.

Hell Fast
Each word makes a checklist (called a query vector) describing the characteristics of words it is looking for. Each word also makes a checklist (called a key vector) describing its own characteristics.

Hell Fast
Attention is all you have 1. You have limited attention so you have to ignore the things that is familiar 2. You have to be reluctant to things that contradict your familiar because it cost you too much 3. You want to add new information as curiosity




Hell Fast
But doing homeworks is an another story.

Hell Fast
That's not the same. It's encouraged to use AI in publishing papers. The scientists who don't use AI will be outdated.

Hell Fast
No. It has never been writed before.

Hell Fast
仅仅两天,已经在substack上“写”了25个文章了 AI is not the copilot, he is the ghostwriter.

Hell Fast
刚起来的时候有两个值得写下来的idea,手机没电了,没记录,现在忘了,就这么稍纵即逝

Hell Fast
写作就是对抗遗忘,防止反刍

Hell Fast
1994年 1. 上幼儿班 2. 动画城开播 3. 三国演义 4. 年轮 5. 东方小故事 6. 包青天 1. 肖申克救赎 2. 低俗小说 3. 老友记 4. 阿甘正传 5. 狮子王 6. 这个杀手不太冷 7. 雅虎 8. 亚马逊 9. OJ Simpson 10. Playstation 11. Netscape / Internet 12. ER 13. 卢旺达 14. 世界杯 15. 乔丹退役
1 replies
Hell Fast
Chauvet Cave Fermat's Last Theorem

Hell Fast
1. The Problem Solver, either solving or preparing for solving 2. The Weak Mind, bias over variance, efficiency over accuracy 3. Thieves and Slaves, how to cooperate with other people 4. The Show Chaser, what drives people to unintentional place 5. The Only One Living, how serious is death
1 replies
Hell Fast
The Only One Alive

Hell Fast
如果七万字,1500字左右,需要50篇文章 如果十万字,1000字左右,需要100篇文章 that's the range

Hell Fast
用了不到半个小时,开发了一个简单的 ai chatbot,就是简单啊

Hell Fast
1. 财富,钱是怎么来的,产品,设计 2. 认知,主要还是association 3. 共识,政治合作,科学进步,反制 4. 虚荣,节目效果,情绪价值 5. 本质,科学视角,唯一存活,局限


Hell Fast
如果ai可以这么精准的找到信息,好多博客就没有浏览量了 相当于科普2*4=8

Hell Fast
"A minimalist and elegant representation of []. The image features a clean and simple background, such as a soft gradient transitioning from white to light []. The overall atmosphere is [], focusing on clarity and visual appeal."

Hell Fast
我要以最快的速度,结束之前很多的思考,ai的浪潮要来了,我不能允许自己错过,我要把全部的精力,投入到这个新的浪潮之中

Hell Fast
另外一件事,其实就是人还是特别想让自己舒服,情感上舒服,他可以为此付出很大的代价比如说交流,他并不想获取知识,而是想让自己觉得舒服,想让别人知道自己很厉害,很有水平,但他并不想让自己提高,其次就是人尝试在各种场合向去表现自己去让自己思考的问题变得简单,即使这样是错的

Hell Fast
现在我越来越意识到人其实就是想得到一个简单的答案,他不想进行复杂的思考,其实个问题本质是复杂的,是需要一个分布或者是一个spectrum的事情,但是人们就想知道简单的¥2的好坏的一个结论

Hell Fast
大家之所以对抽象的讨论难以接受,是因为对实例的接触太少 也就是说,案例见得少,大脑本身也不能对特征进行分析,甚至大脑还没来得及区分或者明确抽象概念,因为抽象概念可能有多种可能性,只有对几个实例进行了cv后,才有可能 与此同时,因为遇到的次数少,大脑也记不住,但你需要更新知识的时候,你根本没有足够的样本来处理这些信息

Hell Fast
有两件事让我没办法有太多的朋友 1. 我的注意力 精力 和 情感分配是有限的 2. 值得成为朋友的人是有限的

Hell Fast
I have friends and I have NPCs For NPCs, I only interact with them to progress in game. For friends, I care their feelings and get warm feelings from them.

Hell Fast
今天让 o1 做一道新出来的 cf 题,很简单 given a, construct an array b such that for every index i, the element a[i] is one of the modes of the subarray [b[1], b[2], ..., b[i]]. 但是 it fails to find a neet way to solve 就是让every 1 to n都出现一遍就行


Hell Fast
来一个新的 threads based on this video tutorial dataset is shakespeare length of dataset in characters: 1115394 length of words 202651 length of paragraphs 32777 www.youtube.com/watch?v=kCc8...

Hell Fast
高智商的人,虽然不容易bias,但容易 variance 高 High intelligence can come with the tendency to overthink situations, analyzing emotions and outcomes in exhaustive detail. This can amplify emotional responses to perceived slights, challenges, or uncertainties.

Hell Fast
Bias 就是 想太少 想的 flexible 不 flexible 的问题 如果 bias 高,就是 makes strong assumptions about the data and fails to capture its complexity oversimplification 但是这么传播效率高,容易理解 efficiency Variance 就是 想太多,太敏感 就是想更加accuracy,但是没办法accurate

Hell Fast
解难题,尽全力

Hell Fast
讲真,只要你多思考几下,压根不用那么复杂 上学的时候要求太高了,以为日常也要高水准输出,其实不用 多搞搞二八原则,然后疯狂输出就对了

Hell Fast
“If we’ve been bamboozled long enough, we tend to reject any evidence of the bamboozle. We’re no longer interested in finding out the truth. The bamboozle has captured us. It’s simply too painful to acknowledge, even to ourselves, that we’ve been taken.”

Hell Fast
I want to be smarter not more famous I want to achieve more not more power

Hell Fast
举两个实际的例子,也就是说最开始的计算机,它的设计功能应该是进行复杂的乘法或者是矩阵的运算,因为这些是人类不擅长或者是希望被替代的功能,而现代人工智能比如说ChatGPT,他希望解决的是拥有比人类更多的知识储备,而不是说能够像人类一样经过很少的几个实例,就能抽象出复杂的规律

Hell Fast
当然,在计算机科学的领域,尤其是人工智能也存在着一种实用性和底层性的平衡也就是说,如果一项技术使用性没那么强,那么他就很难获得足够的吸引力与此同时,我们使用计算机就是想解决人类解决不了的问题或者解决起来困难的问题,而不是那些人类好解决的问题的话,我们才更有销售的必要,但是呢,如果说我们让计算机总是解决我们解决的问题,那就离人类智慧的模拟,就我们更想制造一个更接近于人的智能,但是它的使用性看起来在初期是没那么高的