AI 海啸袭来:重新审视“高地”与“洪泛区”的生存法则
当前关于 AI 泡沫(AI Bubble: 资本过度涌入导致技术估值与实际产出严重脱节的现象)何时破裂的讨论不绝于耳。然而,无论泡沫是否会破裂,一个无可争议的客观事实是:仅在去年,就有高达 6500 亿美元的资金涌入 AI 领域,与此同时,全球有 110 万个工作岗位消失了。这并非暂时的周期波动,而是一场正在逼近的巨型海啸。当海啸即将来临时,海岸线的水位会突然退去,站在沙滩上的人看着自己干燥的脚踝,可能会觉得一切如常。但在千里之外,高达数百米的巨浪正以每小时 500 公里的速度奔袭而来。这正是我们当前所处的“退水期”。
Google DeepMind 的联合创始人 Shane Legg 曾一针见血地指出:“如果你的工作 100% 可以在电脑上完成,那么你正在与成本仅需几美分的 AI 竞争。”随着这场 AI 海啸的迫近,所有人都将被划分到两个截然不同的区域:一个是占人口 1% 的高地(High Ground),他们能够提前布局并获取巨额财富,无论市场崩溃与否都能胜出;另一个则是占人口 99% 的洪泛区(Flood Zone),大多数人将在此处面临失业和贬值的危机。要在这个时代生存,首先必须具备危机感,识别自己是否已经处于被淹没的边缘。
Original English Source
The AI bubble is about to burst. This is what many in Silicon Valley are saying. Maybe true, maybe not. But here's what is true. Last year, $650 billion went into AI and 1.1 million jobs disappeared at the same same time. I've seen cycles like this before. Real estate, crypto, and I get to sit in boardrooms deciding which AI projects get millions in funding and which jobs are going to be at stake. The story is always the same. The smart ones, you know, the top 1% position themselves early and win whether the crash comes or not, while the rest suffer. So, in this video, I want to share that exact playbook so you can stay ahead, become harder to replace, and futureproof your career before it's too late. The first thing you need to understand, you have less time than you think. Imagine standing on a beach. The ocean suddenly pulls back. You look down at your ankles and think, "Well, I'm dry. Looks perfectly safe." But a thousand miles away, sensors show a wall of water racing toward shore at 500 m an hour. That's what happens in every single tsunami. And that receding water moment is exactly where we are with AI right now. Shane Le, co-founder of Google's Deep Mind, puts it very bluntly. He says, "If your job is 100% done on a laptop, you're competing with AI that costs pennies." As this AI tsunami reaches us, it will create two zones for all of us. The high ground where the 1% will survive and capture millions in wealth and success and the flood zone where perhaps 99% of the people will end up standing. Which brings us to the second thing you need to know. How to recognize if you are standing in the flood zone right now. I was in a meeting watching an AI product demo last month. beautifully designed, lighting fast responses, cutting edge AI at the back end. But when I asked, "How many customers use this daily?" Silence. The response was, "Well, we're not ready for prime time yet." But we're working on that right now. And I nodded. I smiled, said nothing. But the translation, a million dollar science project, and that's the Flood Zone reality. Great demos, zero customers, zero cash.
RAIL 评估模型:诊断你的项目是否属于“百万美元的科学实验”
许多 AI 项目在演示时看起来无比惊艳:精美的界面、瞬时的响应以及最前沿的后台模型。然而,一旦深入探究其核心业务指标,往往会陷入尴尬的沉默。这其实是典型的“百万美元科学实验”,其残酷的现实是:大笔烧钱,却拥有零客户与零现金流。为了帮助个人和企业快速诊断是否身处洪泛区,我们可以引入 RAIL 评估模型(RAIL Framework: 一种通过收入、交付速度、市场渗透和持续学习四个维度来评估 AI 项目商业可行性的分析框架)。这个模型包含四个核心问题,仅需 50 秒即可完成评估:
- R 代表收入 (Revenue):你的 AI 是否有真正的付费客户在使用?如果回答依然是“我们仍在试点”或“未来会有的”,这就是一个危险信号。如果你正在领导一个没有清晰收入路径的项目,那么你将是裁员名单上的首选。
- A 代表加速 (Acceleration):你能在两周内向客户交付真正可用的东西吗?如果不能,摩擦力过大就会让竞争对手抢占先机。通过 AI 消除流程中的摩擦力,实现速度和效率的显著提升,是摆脱洪泛区的关键。
- I 代表市场渗透 (In-market):如果你做的是内部工具,它是否已经交付到真实用户手中并开始收集实际数据?仅仅留在测试环境中的工具只是个“业余爱好”,算不上商业资产。
- L 代表持续学习 (Learning):你是否在根据用户的实际反馈进行迭代?这是最核心的瓶颈,因为在现实中观察用户使用产品一周,学到的东西远比六个月的内部测试要多。
例如辅助编程工具 Cursor,他们之所以能成功,就在于他们以极高的速度每天迭代并发布新功能,在用户的实际使用中高频学习。如果你的项目在这四个问题中有两个以上的回答为“否”,那么它就面临着巨大的生存危机。
Original English Source
So, how do you know if your company is in the flood zone? You can apply a framework I call rail. R A I L. Just four questions, takes about 50 seconds. R is revenue. Is your AI used by real paying customers? And if the answer is, well, we're still piloting or eventually we'll get there, that's a red flag. If you are leading a project with no clear revenue path, you're going to be the first one on the chopping block, no revenue, no boat. A is acceleration. Can you deliver something real to customers in 2 weeks? Because if you can't, somebody else will for sure. So figure out where the friction is and how can AI remove it. This is your efficiency play. If you can't show real improvement in speed and efficiency from AI, it's a flood zone. I stands for inmarket. So it's possible that you might be working on an internal tool. Now that's not going to generate cash directly, but that's fine. But is it in the hands of your real users? Is it gathering real data? If it's just sitting on a shelf in a test environment, it's not a business asset. It's a hobby. And finally, L is learning. Are you learning from real customer use? This fourth one is a dealbreaker because one week of watching customers break your product teaches you more than 6 months of internal testing. Also, the pace of AI is so feverish right now that if you are not constantly learning, you're already losing. And I've spoken about companies like cursor before. They are an AI based product that helps you write code and they refine their product on a daily basis launching a new feature every single day. That is the velocity of innovation that we all have to get used to. I'm still getting used to it. If your answer is no to at least two out of those four questions, the company you're building or working for might be in the flood zone. We have to ask the most important question, right? How do you get out of the flood zone? How do you get to the high ground?
AI 时代的三层财富地图:寻找属于个人的商业切入点
想要逃离洪泛区并登上高地,你需要一张清晰的“财富地图”。当前的 AI 产业结构可以被划分为三个层次,而金钱的流向和商业壁垒在不同层次有着本质的区别:
- 第一层:基础设施层 (Infrastructure):由芯片制造商、云服务提供商以及能源供应商组成。这是由 Nvidia 等巨头垄断的庞大资本游戏。除非拥有数百亿美元资金或主权基金支持,否则普通玩家根本无法涉足。
- 第二层:前沿模型层 (Frontier Models):包括开发 ChatGPT、Claude、Gemini 的大模型公司。这一层正处于极速整合中,由于技术同质化,价格战极为惨烈,过去 18 个月中每千个 Token 的成本下降了 98%。同时,开源模型已能达到闭源模型 80% 的效果,成本却大幅降低。在这里创业就如同在今天重新做一家互联网搜索引擎公司,毫无胜算。
- 第三层:应用与服务层 (Apps and Services):这是捕获 AI 商业价值的最佳切入点。如果个人或企业想自主创业,应当专注于此。这一层又可以细分为三个方向:
- 水平 AI 应用 (Horizontal AI Apps):如用于演示制作的 Gamma,以及 AI 搜索的 Perplexity。这类应用面向大众,但竞争极其依赖获客规模与分发渠道(Distribution Battle)。
- 垂直 AI 应用 (Vertical AI Apps):如针对开发者的 Cursor 或是医疗领域的 Open Evidence。深耕特定行业壁垒并占领利基市场(Niche Market),拥有更高的用户粘性。
- AI 服务与集成 (AI Services & Integration):帮助传统企业将 AI 无缝融入其工作流中。像 Scale AI 证明了“人机协同”(Human-in-the-loop)是处理粗糙数据、解决模型局限性的极佳商机。同时,还有许多集成咨询工作室通过为企业提供集成方案,收取每月高达 5 万美元的服务费。他们无需开发任何底层软件,只需作为 AI 落地集成专家协助企业推进数字化即可。
Original English Source
You get there with the right map. The map that 99% of the people will never see. The AI world today looks so confusing and chaotic from the outside. Orchestration layer, vector databases, tensor chips, things that didn't exist 10 years ago. But here's what you should do. Ignore all that noise. All you need is a simple map that shows you the three layers where the money is made. Layer number one, infrastructure. And this is where the fabs and chip makers and cloud and energy providers live. This is a game of massive capital. Unless you're Nvidia or have sovereign fund backing, it's hard to win here. Layer 2 is the frontier models. Chad GPT, Claude, Gemini, XAI, and all the tooling companies around those models. This layer is consolidating really fast. Everyone's copying each other's features. Price per token has dropped by 98% in 18 months, and also open-source models are closing the gap. They're 80% as good, but way cheaper. So, think about it. You wouldn't start an internet search company today, would you? Why start a meto model company? Makes no sense. Now think about the dotcom era. Remember the companies that laid fiber optic cables? Most of them went bankrupt. The ones who build apps on top of them like Amazon, YouTube, PayPal, they won. And that brings us to layer three, which is about apps and services. This is the interface, the solution. This layer is where the value is captured. If you're starting something on your own, this is where you should focus. Now, there are three ways to think about this layer. Horizontal AI apps, vertical AI apps, and AI services. Horizontal AI apps work for everyone. So, think of apps like Gamma for presentation or perplexity for AI based search. Huge market, but it's a distribution battle. You need millions of users to become profitable. Vertical apps on the other hand go deep. Think like cursor for developers or open evidence for medicine. Smaller market but you own your niche. And finally AI services and they're the implementation layer. Companies like scale AI has proven that humans in the loop is a massive business opportunity because they handle the messy data work that models can't do yet. The second big services opportunity here is integration. And I know boutique firms right now that charge $50,000 a month to help traditional companies plug AI into their workflows. And they're not building software. They're just implementing it. They're the experts who charge you to set it up. So that's the map. Layer 1 and two are where the giants are fighting. Layer three is where you can build something amazing. Here's what most people miss. Knowing the map isn't enough, right? You need to know how to avoid the potholes when you're moving to the high ground.
避坑指南与个人转型:从“脚本化执行”迈向“战略性判断”
虽然 AI 带来的行业震荡不可避免,但全面的人类岗位替代尚未彻底到来。目前像微软、亚马逊、Meta、谷歌等科技巨头,主要是在削减非 AI 相关的传统岗位,从而将节省下来的资金用于争夺高稀缺度的 AI 人才。在这样的背景下,如果你的工作只是在“维持现状”,那么你必然处于洪泛区。
要登上高地,必须立即做出转型调整。AI 会首先将任何工作中脚本化(Scripted: 指那些规则明确、可重复且程式化的任务)的部分自动化,而人类生存的唯一根基在于掌控战略性(Strategic: 依赖直觉、复杂推理与人际信任的判断性工作)的部分。个人转型的三步行动方案如下:
- 审计你的工作:剖析你日常工作中有多少时间属于脚本化,有多少时间属于战略性。脚本化工作包括数据处理、会议纪要、常规内部邮件撰写、样板代码编写等;战略性工作则是解决前所未见的问题、建立客户关系、设计系统架构以及捕获 AI 忽略的边缘情况。
- 自我革新:在你的老板或公司用 AI 替代你之前,主动使用 AI 把自己从脚本化的琐碎事务中解放出来。
- 时间再投资:将释放出的时间投入到提升你在战略维度的核心竞争力上。这听起来有些反直觉,但在未来,有价值的工作往往更偏向“白日梦”——通过刻意切断连接、散步和深度思考,在大脑中构建 AI 无法通过信息堆砌得出的直觉与洞察。
Original English Source
That's why number four is about avoiding the potholes when you get to the high ground. The job apocalypse from AI may be inevitable, but it's not here yet. Sure, hiring has slowed down, but companies are still hiring at a slower pace. And the companies that are laying off people first are also the earliest adopters of AI. Microsoft, Amazon, Meta, Google, others, and they're cutting nonAI roles to fund AI talent. So in this world, if your job is to maintain the status quo, you are unfortunately in the flood zone. So to get to the high ground, you have to move. AI will automate the scripted parts of any job. Your survival depends on owning the strategic parts. It doesn't matter if you're a software engineer or a consultant or a data annotator or accountant or analyst. The pattern is going to be the same. Scripted work gets automated. Strategic judgment stays human. Here's your three-step action plan. First step, audit your job. How much of your day is scripted versus strategic? Scripted work is like the mind-numbing repetitive activities. processing data, summarizing meetings, drafting standard internal emails, writing boilerplate code, whatever. Strategic work is about solving problems no one has seen before, building client relationships, designing system architecture, catching edge cases that AI sometimes misses. So, audit your work and divide it into these two categories, scripted and strategic. Step number two, fire yourself from the scripted parts. Automate all of it before your boss does it for you. And step number three, reinvest that time in developing your human edge on the strategic side. Now, this sounds counterintuitive somehow, but I work less on execution now and more on daydreaming. I actively disconnect, stare out the window, take long walks, and from the outside, it may not look like work at all, but this is deep work because deep thinking requires silence, not scrolling. It requires imagination, not information.
黄金生存法则“3R”:以确定的内核应对不确定的 AI 巨浪
培养高效的习惯若无正确的底层心智模式支撑,依然无法让你立于不败之地。在这个多变的时代,能够在 AI 海啸中登上高地的人,都共享着 3R 心智模型:
- 第一个 R 是严谨 (Rigor):指专注于领域底层的深度掌控与边缘情况的精细处理。以医疗 AI 领域的决策平台 Open Evidence 为例,当无数平庸的医疗 AI 初创公司迅速倒闭时,它却凭借着严密的同行评议、详尽的研究数据和对医学边缘情况的反复调试,估值达 35 亿美元,赢得了全美 40% 医生的日常信赖。哈佛与 BCG 的联合研究也证实:在 GPT-4 的辅助下,对业务领域有深度、严谨理解的人,其产出质量能提升 40%,而缺乏深度的人仅能提升 12%。
- 第二个 R 是关系 (Relationship):在算法可以任意生成万物的时代,人类最难被复制的资产就是“生成信任的能力”。在职场中,每一次重大的职业跃迁(如进入华尔街、转型科技行业、出任 CEO 或获得董事会席位)往往都源于深刻的人际联结,而非简历上的文字。倾听他人、建立深层信任才是最坚固的职业护城河。
- 第三个 R 是韧性 (Resilience):指面对失败与变化时,快速迭代并长期坚守的适应力。纵观所有伟大的回归故事——被自己创立的公司驱逐后重返苹果的乔布斯、历经票房惨淡后重新成为宝莱坞之王的沙鲁克·汗、历经 27 年铁窗生涯最终当选南非总统的曼德拉。他们失败了,但选择继续留在牌局中,通过转型与适应生存下来。
正如巨浪拍击礁石时,小波浪会因恐惧岩石而大喊:“我们要毁灭了。”但海洋却微笑地告诉它:“你以为自己只是一朵浪花,其实你就是水,你就是海洋。”在外部世界的巨变中,职位、技术、市场都是转瞬即逝的“浪花”,而你的严谨、关系与韧性,才是你坚不可摧的“水之本质”。
Original English Source
That is where you can connect the dots that others cannot see. Cultivating these right habits alone won't save you. Not without the right mindset. The final framework you should know is the three Rs. You know, the top 1% who will make it to the high ground will all share three things in common. The three Rs, rigor, relationship, and resilience. There are hundreds of AI startups in healthcare. They have come and gone. But open evidence is worth $3.5 billion right now, and it's used by 40% of US doctors daily. Why rigor? They fine-tuned their AI with such meticulous peer-reviewed research and human feedback. Doctors just love them. They evaluated every edge case that would break most other startups. They understood their domain and the rigor it takes to succeed in it. And the data backs this up. BCG and Harvard ran a study with consultants using GPT4. Those with a deep understanding of their domain got 40% better results. Those without only 12%. So that's the first thing you need to stay on high ground. Rigor, deep understanding of your domain, your process, your data, not surface level knowledge, but deep mastery of what you do. Second are relationships. I learned this simple truth while working at a company in New York about a decade ago. It was long hours, brutal pace. So I was there till late night most of the time and I would always keep my office door open in the evening. My colleagues would stop by and we would have a chat. One of my colleagues would often stop by and we would talk for a while about lots of things about his career, his fears, what he really wanted to do. And years later, he became the CEO of a $150 million company. And he told me that those conversations changed his mindset and trajectory. Not because I said something profound, I didn't, but because someone actually was listening, he felt seen and we were able to build a strong relationship. Every major move in my career, Wall Street, tech, CEO role, board seats, happened because of my relationships, not my resume. In a world where AI can generate anything, you have to generate trust. And the final R is resilience. If you think about the stories that inspire us, there is always a pattern. Steve Jobs fired from the company he founded, came back 12 years later and saved it. Shah Ruk Khan a long series of flops. Then he comes back and becomes the king of Bollywood again. Mandela 27 years in prison and emerging to become the president of South Africa. We all love the comeback stories. The ones who fail but still endure. pivoting, adapting, surviving, staying in the game long after everyone else has quit. Resilience and adaptability are etched in our human DNA. When I was young and going through massive challenges in my life, my teacher told me a short story. He said, "Imagine a small wave racing toward the shore. The wave sees the rocks and starts panicking. It screams into the ocean. Why aren't you afraid? Look ahead. We're going to crash. It's over. And the ocean looks back and smiles and says, "You still don't understand, do you?" And says, "You think you're a wave. You're water. You are the ocean. Most external things in our lives are waves. titles, roles, markets, technologies, they rise, they fall, but your rigor, your relationships, your resilience, that's the water. The form changes, the substance remains the same. Let the tsunami come and the waves will crash. In the long run, you'll survive and thrive. So, I hope that you go find your high ground. There's still time. If you like this video, please subscribe so others may find us too.
📌 文中提及的人物和组织
公司/组织: Google, Nvidia, Scale AI, Open Evidence
产品/模型: GPT-4, Cursor, Gamma, Perplexity