Workflow
奥卡姆剃刀
icon
Search documents
高手打的是明牌
3 6 Ke· 2025-07-04 06:56
Group 1 - The article emphasizes the importance of "playing with a clear hand" in investment, suggesting that successful projects are often obtained through transparent and market-driven methods rather than relying on personal connections [1][3][4] - It discusses the current trend in AI investments, highlighting that investing in established companies like Nvidia may be more reliable than betting on startups [5][6] - The concept of "playing with a clear hand" is defined as making decisions based on common sense, transparency, and straightforwardness, which leads to higher success probabilities [9][11] Group 2 - The article outlines several reasons why experienced investors prefer "playing with a clear hand," including a belief in probability over luck, the importance of replicable strategies, and the value of time [12][13][18] - It notes that while "dark cards" (hidden strategies) may yield occasional success, they are difficult to replicate and sustain over time [15][16] - The article also highlights that the pressure to conform to conventional wisdom can deter individuals from making straightforward investment choices, as they may seek to prove their uniqueness [19][24][26] Group 3 - The article concludes with a philosophical perspective on "playing with a clear hand," suggesting that it represents a return to simplicity and common sense in both investment and daily life [30][33][38] - It posits that the most successful individuals often embrace basic principles and consistent efforts, leading to long-term benefits [32][36] - The narrative encourages individuals to recognize the value of straightforward approaches, which can lead to extraordinary outcomes despite their simplicity [31][35][39]
用AI两年半,我常用到的12个思维模型
Hu Xiu· 2025-06-16 06:40
2022年12月,第一次试用ChatGPT,一下子着迷。 于是,两个月后,几乎不怎么写公众号的我,因为ChatGPT,又开始了一段时间的公众号创作。 2023年2月10日,写了第一篇关于ChatGPT的思考:《关于ChatGPT的20条思考,尤其是赚钱的机会》。 现在回看这篇文章,还挺让人回味的。 基于这些基本的初步认知,我开启了AI学习与公司转型之旅,通过两年多的学习与探索,无论是之前还保留的业务,还是目前正在进行中的AI项目,对 我来说,因为AI,目前的状态是创业六年中最好的时候。 正是有了这两年多的各种探索和实践,才能和向阳在很短的时间内原创设计出一套AI课程理论体系并产品化。 因此,借着这个机会,也给大家分享和总结下,用AI两年半,我经常用到的这12个思维模型,它们对我的学习与创业帮助很大。 它们刚好对应: 发现问题 → 拆解问题 → 验证/迭代三个阶段。 一、发现问题 大部分的AI想法试验失败,不是技术不行,而是方向和问题就错了。 可以想想:我们把多少时间、金钱、精力,浪费在了错误的问题上? 1. 乔哈里窗 | 我知道 | AI 知道 | મી પં | | --- | --- | --- | | ...
如何理解李想说理想重新发明了增程?
理想TOP2· 2025-06-12 03:50
2023年7月1日理想八周年的时候李想说:"我们努力解决电动车长途充电难的问题,重新发明了大电池 的增程电动车。" 李想2019年9月说: 理想ONE上的增程电动技术的算法灵感源于苹果的Fusion Drive, SSD安静速度快,但是价格贵容量小,那就用算法把SSD和HD的优点融合起来。电动车的电池同样 安静速度快,但是价格贵、重量大、充电麻烦。我们预留大量的电池容量就开启增程,就是为了让体 验和性能的需求都发生在电池上,让续航的需求发生在燃油增程上,汇集电池和燃油的优点。不像传 统的增程:燃油仅仅是用来续命的。整个技术的架构和算法难度比传统的增程式高了几个量级,把功 率、能耗、NVH、预测四个层面汇集在一套算法系统上,把增程做出来容易,体验做好了还是极具 挑战的,我们一直闷着头做了四年。 群友1:难怪我解除增程的时候一点都不觉得有问题,计算机里面很多类似技术。就是 cache,缓存。每层存储都有cache,从CPU到硬盘,可能有十几级cache。 群友2:不大一样,可以理解为以SSD为主还是HDD为主,SSD是HDD的缓存,还是HDD 是SSD的后备,两种思路。 群友1:cache是有冗余副本,fusio ...