最近发现效率有待提升，遇事给自己紧凑安排，包括timebox, 记录时间。
但是这种安排固然很有效的把效率提升上去，但是问题出在午睡时，不太容易入睡，因为比较紧。脑子处在思考状态
所以需要各种放松的可能性。
https://chatgpt.com/c/676e3f7b-1ad0-8010-963e-e37e102910e0
1. 吃零食 （有用）
2. 看电影 （有用）
3. 旅游 （有用-环境让你忘记）
4. 禅定 （会太专注，放松应该有用）
5. 泡脚 （应该有用）
6. 运动走走，骑自行车（应该有用）
7. 做数学 （应该有用）
8. 写日志（可能有用）
9. 5-4-3-2-1 Grounding Method (应该有用）
---
8. 香味 （得试试）
9. 做算法 （得试试）
--
放松的感觉是肌肉松弛下来了，开心了，然后入睡了。
--
Absolutely—sometimes you just need quick, no-fuss ways to unwind without adding another routine to your already tight schedule. Here are a few simple “in-between” strategies:
Micro-Breaks
Mini Breathing Exercises
Mindful Moments in Everyday Tasks
Gentle Movement “On the Go”
One-Song Reset
Visual Break
Mental Check-Ins
Bottom line: You don’t need a long routine or extra tools to take the edge off. Just sprinkle these mini strategies into your day—like hitting a “pause” button for a few seconds—and you’ll find small but effective moments of relaxation within a tight schedule.
Here are a few suggestions that might help you relax and improve your sleep quality when you have a tight schedule:
Wind-Down Routine
Breathing Exercises
Gentle Stretching or Yoga
Mindful Meditation
Short Afternoon Rest
Aromatherapy or Herbal Tea
Journaling
Light Exercise and Fresh Air
Seek Professional Help If Needed
Experiment with these suggestions, and find which combination works best for you. By creating a calming environment and routine, you’ll give your mind and body the signals they need to unwind and get better rest.
我感觉先不要动笔，先做Survey
- 写一个程序 such that 给与一个 paper 他能download paper 到指定folder
- 然后帮忙summarize 主要content, 来update 一个markdown, 需要用什么benchmark 来证明state of the arts
- 然后看看你，或你团队，或你朋友 可以给与帮助。
从哪里download paper
- 各大会议网站
- 朋友网站
怎么衡量
- effort
- impact
- sustanability
怎么增加中奖概率
- 多投
---
但是究根结底，我觉得需要围绕text-to-sql 运行，这样能满足effort impact 和 sustainable
--
ok 这么定下来后
其实后续就是要怎么把这个tool 构建起来
--
1. 首先需要把要跑的data 全部下下来
2. 然后购建components输入，输出结果
3. 然后再往上搭建
--
团队里我要分成几个
- research （拿最新research）
- business （根据错误调）
- engineering/data （工程、数据）
--
1. Look at the big picture then decide on things, just choose one thing rather doing multi-objective optimization for easiness
2
感觉人生就想创业。。有没有别的方法得到同样的快感呢？
当然就是内在修行成就了。
二者比较虚，我该如何去具体化。。那也算另一个门子的创业。
这个需要trigger changes。
天天提醒，天天改变，天天写感言。
这是提醒自己的一个方式。
而且专注一样事情，28天。
不看social media
多meditation
减慢生活步法
让自己坚持一些事情
1. 早睡，少吃，多运动
2. 少思，专一，放松
3.
有权利欲望，该抓手上就抓住。
想要成事，想要有效率，对上面毕恭毕敬很相信，对领导做的事情有领悟力，表面淡、实际很上心，善于抓领导重点，善于数字。
为他想下一步，因为 - 想要成事，想要有效率。
我这里：
project - 已经想好。
其他 - 可以提供想法。
Overall
Polite, logical
What's important for him:
1. Relationship with CB
2. With stakeholders to move project
3. Identify new project (intake)
4.
Yes:
1. Talk with CB (but should be on first day, to show respect)
2. import project (genai council approval)
3. meet with Kevin to schedule his coming
4. chat with directors to build relationship (for long time)
5. the overall scrum master planning with the team
6. tell that we could support the work stream
No:
1. too much on courses
2. too much on project details
和internal stack holder - ibm 和 external stackholder
mrmg review or defends - journal review or defineds
work on project like t2s - startup
看来IBM 和 Startup 是学到最多的，再来是Research。
Research - 思辨方式。
IBM - consultant
Startup - moving things forward
最近老是被一个字占心， 忙
确实忙，会议不停
但忙的本质，是把你的空挡时间都塞满了
占心的应该是修行相关
比如有智三摩地
简单来说就是保持轻松，闲适，开朗，专一，享受的心
A好像就是好例子
时间不被工作占领，你一天也是被其他东西占领的。
工作就是修行。其他东西也是修行。
重要是专一和菩提心。
其实我心里已经有个底该怎么打分
写什么只是铺垫
这个底，是建立在日常合作上
比如BT, 忽然间有很多idea, 以及presentation变得很好
再来呢 对于K, 其实更重要是A, 他向上汇报。
对于K主要说到的是要如何
- 对他好，tech组好
- 回答简短精要，有信心
日常很重要，进步自己很重要
领导 - 你不在，工作继续。随时随地你都为交际作准备。会时不时看slack and email。你错过的是各种会议。
IC - 你不在，工作就停了。由focus mode转成 relax mode，一时间不太会交际。会时不时看slack and email。你错过的就是没有开发。
我觉得 最大的blocker是把某些东西看成恶，去执着。
无善无恶是重点。
一切都是好的，都是来帮助你的，都要感恩，是重点
再来对升职有执着。
尽人事 听天命，得不得无所谓，是重点。
无欲则刚是重点。
再来对做不做得出来执着。
和上面一样。
还有乐观是重点。
---
统而言之，总和言之，一场梦而已。是重点。
你现在做梦。
刚才也在做梦。
--
这会儿，又多了很多未知
如何成事 （找谁要resource, requirement，optimize project path）
这是目前最难部分。。。
前面是对CoE的认识，
现在是对project 的认识。。这更加重要。
phd, 收入什么的，老往脑边挂，不是个好东西，你不要虚荣心。
刚有个perspective很舒服，就是小孩。
我就是个小孩，累了就睡。
startup是我的玩具，起来就玩，phd，收入什么的都是我的游戏中的纸币。。不想玩就睡个觉。
很自然，就和自己小时候上幼儿园一样一样。
很舒服，稚嫩，愉悦。
成功是谁定义的？
自己。
我刚刚推拿非常疼，我定义成人间至乐。最最快乐的。
我定义成功就是现在的状态，不上不下，刚刚好。
这和昨天的无条件正面和信心一个意思
别怕失败，别怕尴尬
这世界除了你，还有谁？
甚至有你吗？
本来就是一场梦，大家都不存在。
在梦中，找出梦中最好的生活方法。
你是梦的主人，自己定义，统统都定义成好的。
大部分都是可更正，但这些是最好的学习机会，
避免失败，避免尴尬，如果是一场考试，你要考100分，这个是极其巨大的准备 （不知道要准备什么，所以准备全部），和有极其巨大的压力（完美的压力）
但是你如果你不怕，那你就知道重要考题是什么(准备曾经失败的部分），和没有太大压力 （失败无所谓）
所以说，保持乐观极为重要。
与其乐观着失败，也不要悲观着成功。
乐观着失败，从学习学习，更容易成功。
悲观着成功，压力无措，更容易失败。
I 型善于从独处得到能量，E型善于从朋友得到能量。
问了一下E型的人，发现方法是吐槽。
要做E型，我感觉，首先要把我看的淡一点，把别人看的重一点。
- 不熟悉的人，少问隐私，多提供干活，多帮忙
- 幽默点
- 体贴
- 反正就是自他互换。
这样良性循环。
--
另外，对于自己的小点错误，或者别人的无心只过，或语言，大大咧咧的带过去，就好了。
如果有那么点负面情绪，站在禅定的维度看着，让它自然消去。
然后帮自己积累点正面情绪。
别为一场人际关系有什么期待，或害怕失去什么，就是王道。
--
就好像你在写程序一样，想想你是怎么在独处中得到能量。
--
人怎么过得更好
1. 更多享受
从财色明食睡，和感官入手 - 但这些都是苦的根源
2. 更多利他
他是谁？实有吗？你做这个为了什么？
3. 梦醒
积极寻求看到真相，把2当成方法。
嗯，感觉毕竟3才是王道。
其他的，都是假命题，没有可能可以更好了，都是来经历需要经历的；得到该得到的，得不到不该得到的。
看着年轻人，羡慕。如果不为了梦醒，也是梦中行梦事，一晃就没了。
1. more organize before give speech (don't start with murmur than give things)
2. less redundant (avoid common sense)
3. motivate and pacify and positive
4. active listening
5. no interject
第三个星期，我基本可以分辨leader 无障碍，第一基本大部分搞懂，第二星期，Harish vs Rahul, Divya vs Milan, Nisarg vs Sumit, 没搞懂，这些cluster无法分辨。
现在对公司各个角落位置大概也清楚了。也对CoE, 自己的team的分工，和合作，有了更多的认识。
大概第三个星期到一个月，心里已经和IBM有个分割了，看到Slack 和 Outlook 和想到公司的情况，都是现公司的。IBM毕竟一个呆了五年的公司。
应该是以一个利他的心态去完成。
这样合乎修行，合乎工作，合乎因果。
利己，没问题，但是不要建立在压力上，要建立在热情上。
建立impact 是利他的一部分
不要通过压制别人，或藏匿信息。
不要担心工作岗位。
脚踏实地，这里的点就是利他
1. 从你的工作开始
2. 再从你的分享开始
利他，也有助于避免被同伴敌视。
有人会说，信息分享，会不会是一种劣势？
我觉得分享信息，最后最大获益者，是自己，这是我的经验。
就是你会记得这份信息，最后变成自己的参考。
别人的话，其实不会太在意这份信息分享。
就算在意了，用了，因为你的引导，他也会把这份信息分享，大部分是个良性循环。
如果不愿意分享，那这种狭窄的心态，过得也挺憋屈的，憋屈的人气质都不太好，不太有利于自我发展。
另外，比如某人积极分享信息，会不会成为一种优势？
至少形象，不是小气巴拉。
因为别人也想利己，所以形象自然提升。
但有没有变成决定性优势 - 这不一定。
有没有可能变成一种快乐，大概率。
变成一种良性因果，绝对的。
时间到了会变成一种优势，绝对的。
---
利他也需基于真诚和自愿的原则，不应该是压力或义务的产物。真正的利他行为源自内心的热情和愿望，而不是外界的强迫。
so for now you need to prepare
1. good resume (make sure your resume is good...bad resume will just get filtered off before you get interview)
- you can send to me, I can help you change it
2. self introduction (HR screening. or first interview)
- you can introduce yourself to me, i can help you hear it
3. Technical (like accounting)
- go glassdoor and see what is typically ask for the role you apply, prepare for common
4. Behavioral
- typically this is like 3rd, 4th interview...use star method to answer it..
I recommend you go by sequence 1, 2, 3, 4
After you get a new job
you can start to learn from job first, and see whether you have opportunity for value increfment (salary, role), not all company and team has such oppty
if yes then you can start to work on the following:
1. Make sure you understand the job well, and delivered concrete result
2. Make sure the result is well communicated (this is where the business English will be used)
The purpose of 1,2 is to get you a good rating..and potential opportunity for promotion
When you move further you will realize that
1. you spend exactly the same or maybe lesser effort (esp. to lower tier job), but you get much better income
2. at the beginning, hard technical work is differentiator, but when you grow further, communication will become more and more important, it will become a differentiator
3. there is always a ceiling for growth...when you go to certain stage..you will reach a plateau
But for now...the ceiling is still far away...
IBM not in office
a. don't have visibility
b. don't have interaction with any senior management
c. not that easy if needs to get like interaction things done
d. don't know the trend in office
In office the good thing is:
Network/Visibility
Get things solved fast and do it impromptu
Easy to observe the trend
More effective at the time before and after (force no sleep)
随时被PIP- 那是一种状态
新工作待命中 - 那是另一种状态
前者，危机感，各种学习，各种找工作
后者，放松感，各种看
想一想，其实最好的状态是两者中间。
让自己做对的事，专一，放松。
就像GM那样，调得刚刚好就是美。
pip - 没法，必须辞职，ESP 也没有；真正pip 辞职时 - 说成熟善于管理pip, 不然有点像背叛
技术 - 还可以
长远计划 - 全靠吹，而且还idealism，不practical （船快沉了(IQP=0, TQP还提到90%)，还着重船内的装饰 (我们重来不给tech 做tech demo)
向上管理 - 可以，全靠听话
向下管理 - 不怎么关心，没有1-1，拉个meeting都挺难
项目管理 - adhoc, 乱糟糟
要求升职 - 打哈拉，只有偏重某些人
--
升职可能 - 非常小
学习/present - 自己想学就学，想present 就present
不过整体还是worklife balance 的，以前老板是想着怎么说，这个就不需要，就是有时候很多project 一起，然后还会乱答应ROI 不高的project.
最近发现自己妄念比较多
是干扰吗？
后来发现是源于对一件事情的执念，当心执行不当。
醒也想，梦也想，还各种延伸的想。
然后妄念，然后变得有点弱了。
每每禅定，空了这些，就觉得无比轻松。
这些算大杂念，空了，就剩下小小的念头。
松，乐，明。
出了定，回到这些，又被绑。
看清楚了也好，有执则懦弱，无欲则刚强。
https://chatgpt.com/c/8288b61e-3022-4460-8df7-b59ec9cb1e79
1. Promotion - Create new business line, with promise of revenue
2. Networking - Share and Be helpful and be Opened (don't just not contribute unless you are invited, leave that to you family)
3. Presentation - Highlight value (and innovation), project financial growth1. 按照时间进行, 然后番茄时间
---
1. 不能吃好吃的
2. 不上social media
3. 不上news
4. 心乱则持经禅定放松
5. 不能提到职位赚钱
6. 专心在工作上
7. 这个月所有东西，都免费，都是为人儿做。
8. 头陀为第一
9. 不午睡
天份 - 起点高，容易受挫
愚公 - 受挫，但resilient, 没有起点报复，一直望前进，凭的是信心
所以信心比天份重要。
想想当年比我天份高的多的去。
但是比我resilient, 一直往前行的，好像就少很多。
我凭的是勤能补拙，和坚持乐观。
于是从胜算不大的马来西亚普通家庭，出来了（一些些）。
现在修行路，也一样。
天份胜算大。
我这种没有天分的，自己给自己加油努力鼓掌。
最近从床上起来，忽然回到了以前的状态，想要在财色名食睡上，享受。
但是，回想一下，不是在联系意清净吗。
于是收了下。
前者，其实短暂快乐，大部分时候都是不可控，和最终是痛苦的。
但后者，你会觉得很可控，心里一直有阵阵的快乐，光明，阵阵的满足感，一直往前进。
这是目前的感觉，远比以前可控，快乐，光明，及稳定向前。
怪不得说受是苦。
大迦叶头陀第一，不让自己在财色名食睡上打滚，他其实是快乐的，解脱了束缚，怎能不快乐。
--
盘点意清净
还是很执
可控就像GM画画一样完成
画中，放松、专一
画完，忘
不可控，就忘忘忘
--
盘点没有意义活动
social media - 从现在开始，不看
松呢？ Reading, Meditation or Deep Breathing, Listening to Music or Podcasts, Walking, Cycling
--
盘点有意义活动
Recommendation System, Anomaly Detection
今天准备CV里的Pharmacy
今天多加一禅
---
Elon musk achieves two things - make world greener, make astronimal travel easier
Start with vision in mind
e.g., if make word greener then
- find out what pollute the world the most
- engineer a replacement solution
- go into the engineering
Ultimately, there is only so many worthwhile visions in mind.
With a vision in mind, you work for the vision, not the money or other things.
That will be long time sustainable and will be more effective overall.
That's why He initialization and my method both novel, And He even easier, but he is rooted from a better perspective, that perspective is more important then implementation.
Choose vision >> implementation.
看着自己的paper, 想法多新奇，但是解决的问题都不太重要
看着这种bert 什么的 wordpiece encoder,想法同样新奇，但是解决的场景确实很重要的。
lesson - 选个重要的场景，花着同样的思维，带来同样的研究快感，但产生impact却大大不同。
从现在望向过去，什么是重要？
其实也难说。。。generative AI 一出来，曾经重要又好用的的比如conditional random field，都被取代了。。
重要其实还有阶段性。
但又一些work比如resnet, 比如He Initialization, 想法简单，但是因为够底层，所以relatively invariant to new model, 因为new model也要用上。
In the fast-paced world of startups, distinguishing groundbreaking work from the ordinary is crucial, mirroring the challenges in academic research. Drawing inspiration from Fei-Fei Li's 2009 advice on crafting impactful research papers in computer vision, startups can glean vital insights to propel their innovations forward.
Much like in academic research, where only a handful of papers make a lasting impression despite thousands being published, startups must aim to solve significant problems that genuinely advance their field. It's not enough to introduce a new product or service—success hinges on addressing pivotal challenges that reshape the market or consumer behavior. This focus on impact rather than novelty ensures that a startup doesn’t just create a splash, but actually changes the course of its industry.
Fei-Fei Li emphasized the importance of producing work that could lead and inspire future studies, rather than merely adding to the noise. For startups, this translates to developing products or services that aren't just incrementally better, but are capable of setting new benchmarks or creating new markets. In an ecosystem flooded with new ventures, those that prioritize depth and transformative potential in their offerings are the ones that carve out a lasting presence.
Echoing Einstein's principle of simplicity, startups should strive to create solutions that are as straightforward as possible yet fully effective. This approach not only makes the product more user-friendly but also enhances its adaptability and scalability. Complexity should only be as deep as necessary to provide a unique value that simpler solutions cannot.
The ability to communicate effectively is as crucial in academia as it is in the startup world. For startups, this means crafting a clear, compelling narrative not only for potential customers but also for investors, partners, and employees. Every communication should reflect the startup’s vision, its understanding of the problem, and the uniqueness of its solution, much like a well-written research paper communicates its thesis and findings.
Fei-Fei Li's counsel to her students to aim for papers that would be remembered for years parallels the startup need for long-term thinking. Startups should not be seduced by short-term gains or superficial metrics of success. Instead, they should measure their impact by their ability to remain relevant and lead in their sector over time.
Just as rigorous experimentation and validation are indispensable in research, startups must also rigorously test their products and business models. This involves not just initial validation but ongoing testing to adapt to changing market dynamics and consumer needs. This continuous loop of feedback and improvement is vital for sustaining innovation and relevance.
By applying the principles outlined by Fei-Fei Li for producing significant research papers, startups can enhance their chances of success in an increasingly competitive and saturated market. The emphasis should be on solving important problems, ensuring simplicity in solutions, communicating effectively, and aiming for lasting impact. These guidelines serve as a beacon for startups aiming not just to exist but to excel and redefine the boundaries of what is possible in their industries.
Several startups in the past have exemplified the principles of solving significant problems, ensuring simplicity, effective communication, rigorous validation, and aiming for lasting impact. Here are a few notable examples:
Significant Problem Solved: Tesla tackled the significant issue of sustainable transportation with its development of electric cars, aiming to reduce reliance on fossil fuels and decrease carbon emissions. Simplicity: Tesla’s innovation lies in its integration of cutting-edge technology with user-friendly interfaces, making electric cars appealing and accessible to a broader audience. Effective Communication: Tesla's branding and Elon Musk's vision of a sustainable future have been clearly communicated, resonating strongly with consumers and investors alike.
Significant Problem Solved: Airbnb addressed the problem of expensive and limited hotel accommodations by creating a platform where people can rent out their homes or rooms, providing a more personalized and often cheaper alternative. Simplicity: The concept behind Airbnb is simple—allow homeowners to generate income from unused space while providing travelers affordable lodging options. Effective Communication: Airbnb has excelled in communicating a sense of community and local experience, which has been central to its marketing strategy.
Significant Problem Solved: Stripe made it significantly easier for online businesses and e-commerce platforms to integrate payment processing, addressing the complexity that developers faced with other payment gateways. Simplicity: The platform is known for its straightforward API and easy integration process, which have become industry benchmarks for payment solutions. Effective Communication: Stripe has effectively communicated its value proposition to developers first, ensuring it became the go-to solution for startups needing reliable and easy-to-implement payment systems.
Significant Problem Solved: Spotify changed the music industry by offering an all-you-can-stream model, solving problems related to illegal downloads and limited access to music. Simplicity: Spotify’s user interface is intuitive, allowing users easy access to a vast library of music with a simple search. Effective Communication: Its model of legal music streaming combined with its ability to personalize playlists and discover new artists has been key to its communication strategy.
Significant Problem Solved: Google revolutionized the way information is accessed by creating a powerful and accurate search engine that quickly sorts through vast amounts of data. Simplicity: Google’s homepage is famously simple, featuring mostly just a search bar, which underscores its user-friendly design. Effective Communication: Google has consistently communicated a clear vision of organizing the world’s information and making it universally accessible and useful.
Each of these startups has demonstrated the importance of addressing key problems with innovative solutions that are simple, well-communicated, and rigorously tested. They not only found immediate success but have also made long-lasting impacts on their respective industries.
