我目前人生最意外也是最幸运的一个选择:学了一个超级水的专业——生物信息学。
虽然我目前大三,站在现在的时间点往回看我是幸运的,可时间线拉长到10年后呢?再回来看我还是幸运的吗?大概率是吧。
讲讲我的原因:
1. ai让技能的壁垒逐渐降低,拥有一项技能带来的收益已是强弩之末,虽说生物信息学本身不能给你带来多少收益,但“不患寡而患不均”,好专业的红利在降级,对于我这种差专业来说某种程度上带来了“均”。
2. 在ai还没有如此强大的前俩年我想明白了很多,曾经我也是做题家,学历至上,成绩就是一切,极其自负且极端。可我很快认清我的自负完全是在自我意淫,我的价值观遭受无数次崩塌。我质疑,我恨,为什么我会有如此的价值观。我怪罪我之前一切的教育,我是无辜的,凭什么要让我独自承担这后果。在无数次来回拉扯后,我选择与世界和解,并依旧热爱这个世界。(这也是我觉得最幸运的原因)
3.技术壁垒还没被ai打破时,为了给自己谋个好前途便到处探索这个世界。最典型的是我接触了web3,这里面的经历让我见识到了更大的世界,更精彩的人生,更宏大的愿景。
4.这个专业很水,我几乎不用学,给了我大量闲暇时间专注于自己生活。
5.上面4点给了我无穷的自信。
1
2
155
25岁前的剧本是每一个优秀学生的模版,28岁在直播间靠cos isa赚钱,我看她直播播了一整天,现在凌晨一点还在直播,大家真的好不容易啊。
愈发觉得一切都是命,人生无常,自己依靠努力能改变的东西少之又少,好残酷的命运。
240
犯规是被人唾弃的行为
战略性犯规是综合全局的聪明行为
整个X上的价值观是轻视学历的
曾经的我也是这么想的:摆脱学历内卷,all in美股,ai,crypto这些领域
对于ai来说:上游基建层普通人踏入门槛极高,下游内卷严重,很难从里面吃到肉。
对于crypto,美股:大陆在可视未来内是不会开放窗口的,说到底终究是个不入流的灰色产业
对于我来说考研的目的并不是提升自己,而是给自己一个合适的缓冲期
我也希望能早点赚钱,可如今通缩下赚钱不是逆势而为吗?
反者道之动,再苟几年,顺势而为
3
331
一般这种活动都以志愿者名义花150雇一天大学生。
当初好奇觉得好玩报名了浪琴的马术比赛,到了后给我安排检票任务,往那一站就是8小时。虽说给了一百五还包一顿午饭,但我当时在上海一天住宿都花300。
另外还有阴阳工资,我旁边那个女生因为以前在他们那干过F1的活动,工资比我多20。
对方要求是三天,我干一天就跑路了,交了60的押金也打水漂,但还好第一天的工资还是结给我了。😂
5
303
3.什么是HyperCore?
交易所每天要处理大量操作:
下单 → 撤单 → 成交 → 平仓 → 清算。
如果这些操作放在一个普通的通用区块链上,速度和成本可能很难达到交易所需要的体验。
所以 Hyperliquid 干脆自己造了一套专门为金融交易优化的系统。
它最核心的部分就是:
HyperCore
它实现了:
CEX 的交易体验+DEX 的链上金融系统
也就是说:
Hyperliquid使用的是链上订单簿:
你
↓
Order
↓
HyperCore
↓
On-chain Order Book
↓
Matching
↓
Trade
交易体验接近 Binance。
但交易状态由链上系统维护。
1
26
Dylan Xu retweeted
上周参加MIT的活动时看到一张挺有意思的图,想跟大家分享一下。
这张图来自@nberpubs的工作论文《Intelligence and the Labor Market》,展示的是不同薪资分位岗位的AI暴露程度。
整体来看,岗位薪资越高,受到AI影响的程度通常也越高。但到了收入最高的前10%左右,这条曲线反而开始下降。
这里需要说明,AI exposure并不等于“被AI取代的概率”。它更准确地反映的是,一个岗位中的工作任务有多大程度可能被AI改变、辅助或自动化。
但这张图还是让我想到一个很现实的问题:
在AI时代,为什么很多高薪知识型岗位会受到更大影响,而真正处于顶尖位置的一小部分人,反而可能更难被替代?
我的理解是,越往上走,一个人的价值就越少来自“把任务完成得更快”,而更多来自几件事:
👉能不能提出真正重要的问题
👉能不能在信息不完整时作出判断
👉能不能协调不同团队、资源和利益关系
👉能不能在关键时刻承担责任
👉能不能让别人相信并愿意跟随你的方向
对高级管理者来说,AI时代真正需要完成的,可能是几个转变。
1️⃣ 从“高端执行者”变成“全局编排者”
AI和团队都可以成为你的执行系统。管理者的价值,不是亲自完成每一项任务,而是定义问题、判断优先级、识别风险,并把分散的产出整合成真正有商业价值的结果。
2️⃣ 建立对组织和人的深度理解
AI拥有大量公开知识,但它并不了解一家公司内部真实的文化、历史、信任关系和利益结构。
它很难知道董事会真正担心什么,客户没有说出口的需求是什么,也未必明白为什么一个理论上正确的方案,在某个组织里就是推不动。这些判断力,只能来自长期参与真实的谈判、协作、冲突和危机处理才能培养出来。
3️⃣ 用AI扩大自己的管理半径
过去,管理者大量时间都花在汇报、沟通、纠错和重复协调上。当AI能够处理更多基础分析、信息整理和流程工作,一个管理者就可以同时看更多项目、理解更复杂的数据,并把精力放在更重要的决策上。
真正善用AI的管理者,未来可能可以覆盖过去一个小型事业部才能管理的业务规模。
4️⃣ 强化那些AI可以增强、但无法真正拥有的能力
判断力、同理心、说服力、信任和领导力,反而会变得更重要。
当越来越多工作被自动化,现实中的人际情感连接反而会变得更加稀缺。团队在面对AI带来的不确定性时,需要的不只是效率工具,也需要有人去解释方向,建立信心,带领团队一起往前走。这是AI不能取代的人文关怀。
所以我越发觉得,AI时代所谓的“不可替代”,并不是你比机器做得更快,而是你能不能做到机器无法独立完成的事:
👉 理解复杂的人,判断模糊的现实,然后带领大家走向一个还没有标准答案的方向。
Last week, I came across a really interesting chart at an MIT event and thought it was worth sharing.
The chart comes from the @nberpubs working paper Artificial Intelligence and the Labor Market. It looks at how exposed different jobs are to AI across wage levels.
The general pattern is quite clear: the higher the salary, the more exposed the job tends to be to AI.
But once you get to roughly the top 10% of earners, the curve starts to come down.
One important clarification: “AI exposure” does not mean the chance of being replaced by AI.
It is more about how much of a job could be changed, supported or automated by AI.
Still, the chart made me think about a very practical question:
Why are so many well-paid knowledge jobs highly exposed to AI, while the small group of people at the very top may actually be harder to replace?
My view is that the higher you go, the less your value comes from simply doing tasks faster.
It increasingly comes from whether you can:
👉 Ask the questions that really matter
👉 Make decisions with incomplete information
👉 Align different teams, resouraces and interests
👉 Take responsibility in critical situations
👉 Earn trust and get people to follow a direction
For senior leaders, I think the AI era requires a few important shifts.
1️⃣ Move from high-level executor to organizational orchestrator
AI and your team can both become part of your execution system.
A leader’s value is not in personally completing every task. It is in defining the right problems, setting priorities, spotting risks and turning scattered outputs into something that creates real business value.
2️⃣ Build a deeper understanding of people and organizations
AI has access to huge amounts of public knowledge, but it does not truly understand the internal culture, history, trust and power dynamics of a specific organization.
It may not know what the board is really worried about.
It may not understand what a client wants but has not said.
And it may not see why a solution that looks completely logical on paper simply will not work inside a particular company.
That kind of judgment comes from years of real negotiation, collaboration, conflict and crisis management.
3️⃣ Use AI to expand your management capacity
A lot of a manager’s time used to be spent on reporting, communication, corrections and repetitive coordination.
As AI takes on more basic analysis, information processing and workflow tasks, leaders can oversee more projects, understand more complex data and spend more time on the decisions that actually matter.
A leader who uses AI well may eventually be able to manage a scale of business that previously required an entire small division.
4️⃣ Strengthen the abilities AI can support, but cannot truly own
Judgment, empathy, persuasion, trust and leadership will become even more important.
As more work becomes automated, genuine human connection may become more valuable, not less.
When teams are dealing with uncertainty caused by AI, they need more than efficiency tools.
They need someone who can explain the direction, build confidence and lead people forward together.
That human side of leadership is still very difficult to automate.
The more I think about it, the more I believe that being “irreplaceable” in the AI era is not about doing the same work faster than a machine.
It is about being able to do what a machine cannot do on its own:
👉 Understand complex people, make sense of an uncertain reality, and lead others toward a future where there is no standard answer.














