Introduction
You've probably wondered: On one hand, NVIDIA's stock keeps hitting new highs, making you feel like you're missing out on a fortune if you don't buy it. On the other hand, some warn that this is a huge bubble, and jumping in now is like buying at the top. So, should ordinary people dive in or stay away?
To answer this question, we need to distinguish between real technology and artificial prices. AI technology itself is revolutionary — like pure water, it's useful. But its price is like soda water, filled with gas — valuation bubbles. Recently, legendary Wall Street investor Jeremy Grantham warned that the AI market is in the final chapter of the 2000 internet bubble. Bank of America's chief investment strategist Michael Hartnett compared NVIDIA's P/E ratio to Cisco's at the peak of the 2000 bubble, and the results are strikingly similar.
Back then, everyone believed Cisco was the cornerstone of the internet, but that didn't stop its stock from plummeting 80% after the bubble burst. On the other hand, Wall Street bulls like Cathie Wood argue that AI, like the invention of electricity and the internet wave, will revolutionize everything around us. Now it can generate movie-level videos, accelerate cancer drug development, and reshape many industries in unprecedented ways.
So who should we believe? To answer this, we can't just focus on current stock prices and news — we need a map, and the best map is hidden in history.
The Survivor Model: Lessons from the 2000 Internet Bubble
Let's rewind to 2000, when the internet bubble burst, countless companies vanished, and countless fortunes were lost. But looking back today, if we sift through all the surviving and dead companies, we find a survivor model that applies to almost all technological revolutions. It's like a pyramid with three layers:
Layer 1: The Foundation Layer ("Sellers of Shovels")
In the gold rush, the first and most certain way to make money was selling shovels and jeans. In 2000, that company was Cisco, which sold routers and switches — the "pipes" of the internet world, without which information couldn't flow.
Characteristics of this layer:
- Closest to the revolution, everyone needs it.
- Most likely to be hyped by capital, leading to valuation bubbles.
Layer 2: The Platform Layer ("Builders")
If the foundation layer provides tools, the platform layer uses these tools to build "super cities" in the new world and set their own rules. Around 2000, Amazon built an "everything store," and Google built an "information entrance." They used new technology to build strong ecosystems connecting countless merchants and users, creating uncrossable moats.
Characteristics of this layer:
- Strong network effects and user stickiness.
- Can turn revolutionary technology into sustainable, stable, and extremely profitable businesses.
Layer 3: The Application Layer ("Gold Diggers")
The top of the pyramid is the most crowded and competitive layer — the application layer, where thousands of "gold diggers" operate. These were the ".com" companies of the past, trying to solve all kinds of problems, from group buying to fresh food delivery.
Some survivors like PayPal became giants, but many more like Pets.com and Webvan burned hundreds of millions of dollars in investment and left nothing behind.
Characteristics of this layer:
- Most innovative, closest to users.
- Most concentrated in bubbles, highest mortality rate — up to 99% of companies may disappear.
Applying the Model to Today's AI Wave
Now let's use this survivor model to look at today's AI landscape:
Foundation Layer: Sellers of Shovels
Undoubtedly, this is NVIDIA, AMD, TSMC, and ASML. Their computing power chips and equipment are the "routers and switches" of the AI era.
Platform Layer: Builders
These are companies like Microsoft, Google, and Adobe. Microsoft injects AI capabilities into Office and Windows through Azure and Copilot; Google integrates Gemini into its search and Android ecosystem; Adobe embeds AI drawing capabilities into Photoshop. They're making us unknowingly inseparable from AI, just like smartphones once did.
Application Layer: Gold Diggers
This layer is countless — video generation tools like Pika and Sora, design tools like Midjourney, search tools like Perplexity, and thousands of writing tools, educational software, and customer service tools. These are today's Pets.com, but among them, there must be the next PayPal.
Three Opportunities for Ordinary People
Now that we understand the map, what should ordinary people do? Based on different risk levels and investment, I've summarized three opportunities:
Opportunity 1: Career Enhancement (Low Risk, Most Suitable for Most People)
The goal is not to switch careers, but to use AI to make your current work 10x more efficient and valuable. In any industry, strive to be the person in the office who best uses AI — this will make you more secure in layoffs and more confident in salary negotiations.
Examples:
- Designers: Master Midjourney and Stable Diffusion to outperform peers in creative efficiency.
- Editors/Operators: Use ChatGPT or Claude to assist with research, data analysis, and content drafting.
- Everyone: Focus on developing skills that AI can't replace, like complex decision-making, deep empathy, and trust-building.
Opportunity 2: Side Hustle & Small Business (Medium Risk)
Suitable for people with professional skills who want to transform or start a side business/small business. The goal is to connect ecosystems and provide services to others for steady income.
Ways to do this:
- Knowledge Sharing: Create content about AI, do corporate training, or become an AI tool reviewer/KOL. Many people are anxious about AI but don't know where to start — if you can make complex things simple, that's valuable.
- Micro-Applications: Develop small tools based on large model APIs to solve specific pain points, like tools for lawyers to write contracts or accountants to review financial statements. These can generate subscription revenue.
The core of this opportunity is to use your expertise to match AI, finding small, real needs that big companies ignore. It may not make you rich, but it has low investment and risk, and can even grow into a sustainable small business.
Opportunity 3: Frontier Innovation (High Risk, High Return)
Only suitable for a very small number of entrepreneurs, venture capitalists, or top technical talents. You can choose to go all-in on AI, with the goal of becoming or investing in the next Amazon or PayPal.
Two approaches:
- Invest in or Join "Killer Apps": Develop your own criteria to identify promising AI companies. Key questions: Does it solve high-frequency, urgent needs? Does it have a data moat? Is its business model clear?
- Focus on Core Technology: If you're a top technical talent, dive into underlying algorithms, model optimization, and computing power architecture — this is the center of the action and the source of future wealth.
Conclusion: It's a Strategy Question, Not a Choice
So, should we dive into the AI wave or stay away? It's not a choice — it's a strategy question. Valuation bubbles will eventually dissipate, like the gas in soda water. But the "water" of the AI technological revolution will remain forever, nourishing our future world.
Our strategy is not to bet on whether the bubble will burst, but to quickly learn to use this new technology and make it work for us. AI can solve more and more problems, but it may never be able to ask a good question. The scarcest resources in the future will not be answers, but high-quality questions, deep insights, and warm empathy. These are our greatest and most fundamental opportunities as humans in this wave.
常见问题
How do I know if AI is in a bubble right now?
You don't need to predict it — and nobody can. Instead, use the survivor model as a lens: Foundation layer stocks (NVIDIA, TSMC) are the most bubble-prone because they're the easiest narrative for capital to hype. Platform layer companies (Microsoft, Google) have real revenue and moats, making them more resilient. Application layer startups have the highest failure rate but also the highest upside. The practical takeaway: don't bet your savings on any single layer. Diversify across layers and time horizons, and focus your personal effort on the layer that matches your skills and risk tolerance.
I'm not a programmer — can I still benefit from the AI wave?
Absolutely. The three opportunities in this guide are designed specifically for this. Opportunity 1 (Career Enhancement) requires zero coding — it's about using existing AI tools to 10x your current work. A marketer using ChatGPT for copywriting, a designer using Midjourney for mockups, a teacher using AI for lesson plans — none of these require programming. Opportunity 2 (Side Hustle) may involve light technical work but can be done with no-code tools. Only Opportunity 3 requires deep technical skills. Most people should focus on Opportunities 1 and 2.
What's the single most important skill to develop in the AI era?
Asking good questions. AI can generate answers, but it can't decide what's worth asking. The ability to frame problems clearly, identify what matters, and iterate on prompts until you get useful output — this is the meta-skill that makes every AI tool more effective. It's also the one skill that becomes more valuable as AI gets better at generating answers. Pair this with deep domain expertise in your field, and you become irreplaceable.
Should I invest in AI stocks now?
This article is not financial advice, but the survivor model offers a framework: if you're investing, distinguish between trading (timing the market) and investing (holding through cycles). For long-term investing, platform-layer companies with real revenue and moats tend to survive bubbles. For short-term trading, understand that foundation-layer stocks are the most volatile — they rise fastest in hype and fall hardest in crashes. The safest approach for most people: dollar-cost average into a diversified tech index fund rather than betting on individual stocks.