AI as the New Internet: Those Who Get It Will Be 10 Years Ahead
Remember When the Internet Looked Like a Toy?
The year was 1995. My uncle was showing me something he called “electronic mail.” We sat at a computer that took up half the desk, and he enthusiastically explained how he sends messages to his colleague in Germany. In a few seconds. Without a postal stamp.
I remember thinking: “Cool, but why would anyone use this when they can just call?”
Ten years later, I worked at a company that built its entire existence on the internet. Companies that dismissed the web as a curiosity for enthusiasts in 1995 either no longer existed or were desperately catching up on lost time.
Today I sit at my computer and watch a similar situation unfold. Except this time it’s about artificial intelligence. And this time I’m paying damn close attention not to make the same mistake as in ninety-five.
Why Most People Still Don’t Get AI
The problem with revolutions is they look boring until they happen. The internet in 1994 was slow, ugly, and complicated. Most websites looked like an enthusiastic high schooler made them (because often they did). People asked: “What would I even use this for?”
AI in 2024 had a similar problem. ChatGPT could write poems and answer questions. Interesting, but revolutionary? More like an amusing toy.
But here’s the catch. The internet didn’t win because websites in 1994 were amazing. It won because it enabled something fundamentally new: connecting people and information in a way that wasn’t possible before. The technology itself was just the means.
AI does something similar, but in a different dimension. It doesn’t connect people with information. It connects people with intelligence. With the capacity to analyze, create, and decide in ways that were previously reserved only for humans.
And this is where the part most people don’t understand begins. It’s not about what AI can do today. It’s about what it will be able to do in five years if current development continues. And if you look at the data, there’s no reason to think it will slow down.
Method: How We Evaluated AI’s Impact on the Labor Market
When I started writing this article, I didn’t want to settle for the usual phrases about how AI will “change everything.” I wanted concrete answers. So I did what I always do when I need to understand a complex problem: I broke it down into parts.
Step 1: Identifying Historical Parallels
First, I looked at previous technological revolutions. The internet, mobile phones, personal computers. In each case, I searched for a pattern: How quickly did people understand the real potential? Who profited first? What did it cost those who came late?
The result was surprisingly consistent. Those who understood the potential within 3-5 years of the technology’s mass adoption gained an advantage that others couldn’t catch up to even in a decade.
Step 2: Analyzing Current Adoption State
Then I looked at data on how companies and individuals currently use AI. I used publicly available surveys, corporate reports, and my own observations from consulting practice.
Finding: Most people use AI as an improved search engine or text editor. Only a small portion (I estimate 5-10%) understands AI as a tool for fundamental change in how work gets done.
Step 3: Projecting Future Development
Finally, I tried to estimate where AI will move in the next 5-10 years. Here I was very conservative. I didn’t assume any breakthroughs or AGI. I simply extrapolated current trends.
Even with a conservative estimate, it turns out that AI in 2031 will be capable of autonomously handling most routine cognitive tasks that currently take up 40-60% of people’s working time.
Those Who Understood the Internet First
To understand what’s happening with AI, let’s look at those who understood the internet before everyone else.
Jeff Bezos worked on Wall Street in 1994. He had a great career and high salary. Then he read a statistic that the web was growing 2,300% annually. He quit his job and founded an online bookstore.
His colleagues thought he’d gone crazy. “Who’s going to buy books through a computer?” they asked. Today Amazon is one of the most valuable companies in the world and Bezos is among the richest people on the planet.
Reed Hastings was paying late fees for overdue DVDs at the rental store in 1997. He got angry and founded Netflix as a service for renting DVDs by mail. That alone wouldn’t have been enough. The key was that from the beginning he counted on the internet changing how people consume content.
When broadband reached critical mass, Netflix was ready. Blockbuster, which refused to buy Netflix for 50 million dollars, went bankrupt.
Marc Andreessen wrote the first popular web browser while still a student. Then he co-founded Netscape and became one of the most influential investors in Silicon Valley. His famous essay “Software Is Eating the World” from 2011 predicted what happened in the following decade.
What do these people have in common? They weren’t geniuses with supernatural ability to see the future. They were people who noticed a trend earlier than others and had the courage to act on it.
And Those Who Missed the Internet
On the other hand, we have stories of those who didn’t understand.
Kodak invented digital photography in 1975. Yes, you read that right. Kodak had the first digital camera. But management was afraid it would destroy their main business with film. So they buried the technology.
Thirty years later, Kodak went bankrupt. Others had taken over digital photography in the meantime.
Nokia dominated the mobile phone market in 2007. When Apple introduced the iPhone, Nokia’s CEO at the time declared it wasn’t competition. People surely don’t want a phone without buttons.
By 2013, Nokia sold its mobile division to Microsoft for a fraction of its former value.
Blockbuster had the chance to buy Netflix for a ridiculous price in 2000. They refused. “Who would want to watch movies over the internet?” The CEO declared that their model with brick-and-mortar stores was unbeatable.
In 2010, Blockbuster declared bankruptcy. Netflix is now worth over 150 billion dollars.
The pattern is clear. Those who rejected or underestimated the new technology paid an enormous price. And the worst part is that at the time, their decisions looked reasonable.
What AI Changes Differently Than the Internet
I have to be honest here. AI isn’t just the “new internet.” It’s something fundamentally different. And this difference is why the impact could be even more dramatic.
The internet changed how we get information and how we communicate. But it didn’t change work itself as much as we think. We still write emails (instead of letters), still make presentations (just in PowerPoint instead of on transparencies), still make calls (just via Zoom instead of telephone).
AI changes the very essence of what “work” means. For the first time in history, we have a tool that can independently perform cognitive tasks. Not just faster or more efficiently. But autonomously.
This is a crucial difference.
When Excel came along, accountants still had to understand accounting. Excel just sped up calculations. When AI accounting comes, it understands accounting itself. And the human? They supervise.
This is a shift that many haven’t understood yet. It’s not about learning to “use AI tools” the same way we learned to use Excel or email. It’s about understanding that the human role is fundamentally changing from executor to supervisor, strategist, and arbiter.
Practical Illustration: A Programmer’s Day in 2026 vs 2020
My friend Martin is a programmer. In 2020, his typical day looked like this:
- Morning: an hour reading documentation for a framework he needs to use
- Late morning: writing code, googling how to do something every 15 minutes
- Afternoon: debugging, finding bugs, reading Stack Overflow
- Evening: refactoring code, writing tests
In 2026, his day looks different:
- Morning: describes to AI what needs to be done, AI generates first version of code
- Late morning: reviews code, adjusts architecture, handles edge cases
- Afternoon: tests with AI assistance, AI automatically fixes common bugs
- Evening: plans next features, consults with product manager
Martin hasn’t stopped being a programmer. But his work has changed. He used to spend 70% of his time writing code. Today he spends 70% of his time thinking about what should be written and reviewing what AI wrote.
Is he faster? Yes. But mainly he does different things. And programmers who haven’t understood this change have a problem. They’re still trying to be fast at writing code. But that’s no longer the key skill.
Three Phases of Adoption for Every Major Technology
When I study the history of technological revolutions, I see a repeating pattern. Every major technology goes through three adoption phases:
Phase 1: Toy (0-3 years) Enthusiasts and experimenters use the technology. Mainstream ignores it or laughs at it. “What’s it good for?” is the most common question.
Phase 2: Tool (3-7 years) The technology becomes a useful tool for specific tasks. Early adopters gain competitive advantage. Mainstream starts paying attention.
Phase 3: Infrastructure (7-15 years) The technology becomes an invisible part of life. Those who don’t use it are at a fundamental disadvantage. Original early adopters have an established position.
The internet in 1993 was a toy. In 1998, a tool. In 2005, infrastructure.
AI passed the toy phase sometime around 2023-2024. We’re now in the middle of the tool phase. The infrastructure phase will come around 2030-2032.
This means we have a window of opportunity. Those who understand AI as a tool now will be ready when it becomes infrastructure.
What Specifically to Do: A Practical Guide
Enough theory. Let’s get to practice. What specifically can you do so you don’t fall behind?
1. Stop Asking “Will AI Replace Me?”
This is the wrong question. A better question is: “How will AI change what I do?” Almost no job will be completely replaced. But almost every job will be transformed.
Your task: Identify 3-5 tasks you do every day. For each one, ask: “Could a person with AI assistance do this task faster and better?“
2. Learn to “Talk to AI”
The most valuable skill of the next decade won’t be programming or data science. It will be the ability to effectively communicate with AI systems. It’s called prompt engineering, but really it’s something deeper.
It’s about the ability to clearly articulate what you want. Break down complex problems into parts. Iteratively refine the assignment. These are skills that will be relevant even when specific AI tools change.
3. Build “AI-Proof” Skills
Some skills AI (for now) won’t replace:
- Strategic thinking and decision-making under uncertainty
- Creative problem-solving without precedent
- Emotional intelligence and working with people
- Ethical reasoning and value-based decisions
- Physical dexterity in unpredictable environments
Invest in these skills. Not because AI could never be good in these areas. But because these will be the last areas where humans remain irreplaceable.
4. Experiment, Experiment, Experiment
The worst thing you can do is wait. Wait until AI “matures.” Wait until it’s clear which tool will win. Wait until your employer tells you what to do.
Every week, try something new. Use AI for a task you’ve done manually until now. Try a different tool. Take on a project you couldn’t handle without AI.
Generative Engine Optimization
As I write this article, I realize the irony of the situation. I’m writing about how AI is changing the world, while knowing that this text will probably be processed by AI systems that will summarize, analyze, and cite it.
This is the new reality. Content is no longer consumed only by people. It’s consumed by algorithms that decide what people will see. And AI-powered search engines and assistants add another layer: they decide not only what people will see, but how they’ll see it.
For authors and creators, this has profound implications. It’s not enough to write for people. You must write in a way that AI systems will correctly understand and interpret. That doesn’t mean writing for robots. It means writing clearly, structurally, and with context.
But there’s something more important. In a world where AI can generate unlimited amounts of average content, the value of an authentic human perspective increases. AI can summarize facts. But it can’t (yet) offer an original perspective based on unique life experience.
My British cat Lilly just came over and lay down on my keyboard. As if to say: “Hey, stop writing about the future and feed me.” AI would never write this sentence. Not because it technically couldn’t. But because it doesn’t have a British cat lying on its keyboard.
This is the meta-skill of the future: understanding when to use AI and when to offer something AI cannot offer. Automation-aware thinking. The ability to see the boundary between what can be automated and what remains the domain of human uniqueness.
Risks and Dark Sides
I would be dishonest if I only wrote about opportunities. AI also brings risks we cannot ignore.
Skill Erosion
When GPS navigates for us, we stop orienting ourselves in space. When spell-checker corrects for us, we stop learning spelling. What happens when AI thinks for us?
This isn’t paranoia. It’s a documented phenomenon. Pilots who rely too much on autopilot lose the ability to fly the plane in crisis situations. Doctors who rely on diagnostic algorithms overlook symptoms the algorithm doesn’t know.
Concentration of Power
AI requires enormous computing capacity and data. This means its development is concentrated in the hands of a few large companies. What happens when these companies gain too much power?
Disinformation on Steroids
AI can generate convincing text, images, and video. In the hands of malicious actors, this means the ability to create disinformation on a previously unthinkable scale.
Existential Risks
And then there are questions that go beyond economy and society. What if we create AI that’s smarter than us? What if we lose control? These aren’t sci-fi scenarios. They’re questions that serious researchers are working on.
I don’t have answers to these questions. Nobody does. But ignoring them would be irresponsible.
Why Right Now Is the Critical Moment
graph TD
A[2020: AI as curiosity] --> B[2023: ChatGPT - mass adoption]
B --> C[2026: AI as work tool]
C --> D[2030: AI as infrastructure]
E[Early adopters] --> F[Competitive advantage]
G[Late adoption] --> H[Catching up on losses]
C --> E
C --> GThis diagram shows where we are. The year 2026 is in the middle of the transition phase. AI is no longer a curiosity, but it’s not yet ubiquitous infrastructure.
This is the window of opportunity. In five years, using AI will be as common as using email. Those who don’t learn it now will be at a disadvantage similar to not knowing how to use the internet in 2010.
Concrete Examples from Different Fields
Let’s look at how AI is changing specific professions:
Lawyers: Due diligence that used to take weeks, AI handles in hours. Lawyers who learned to use AI take more cases. Those who haven’t are losing clients.
Doctors: AI diagnostics already outperforms human specialists in some areas today. But communication with patients and decision-making in ethically complex situations remains key.
Marketers: Content generation, data analysis, personalization - AI handles all of this. Strategic thinking and brand understanding remains the human domain.
Accountants: Routine accounting will be swallowed by automation. But financial consulting and strategic planning requires human judgment.
Teachers: AI can personalize instruction in a way one teacher for 30 students never could. The teacher’s role shifts to mentoring and developing social skills.
Journalists: Sports results and financial reports are written by AI. Investigative journalism and commentary remain the human domain.
The Model of Change: From Executor to Strategist
flowchart LR
subgraph Past
A[Human does work]
end
subgraph Present
B[Human + tools]
end
subgraph Future
C[Human as strategist]
D[AI as executor]
C --> D
end
A --> B --> CThis shift is key to understanding what’s happening. It’s not just that we’ll have better tools. It’s a fundamental change in the role of humans in the work process.
Before: Human performs work, tools help. Now: Human and tools collaborate. Soon: Human directs, AI executes.
Those who understand this change and adapt to it will prosper. Those who try to compete with AI in what AI does better will lose.
Final Reflection
Let’s return to the beginning. In 1995, I didn’t understand why anyone would use email instead of the phone. I made a mistake. I underestimated a technology that changed the world.
Today I see a similar situation. People around me say: “AI is just a toy. It won’t change anything fundamental. My work is too creative/complex/human for AI to replace.”
Maybe they’re right. But when I look at history, I’d bet against them.
It’s not about being a pessimist or optimist. It’s about being a realist. AI is neither savior nor apocalypse. It’s a tool. A powerful tool that will change the rules of the game.
Those who understand this sooner will have a head start. Not because they’re smarter. But because they’ll be prepared.
And those ten years? That’s not a random number. It’s approximately how long it takes for a new technology to become infrastructure. Those who start now will have a decade to build their position. Those who wait will spend a decade catching up.
The choice is yours.
Lilly, my British cat, has meanwhile fallen asleep on the couch. The future of AI clearly doesn’t interest her. Maybe she’s right. Maybe the best strategy is to simply exist, eat kibble, and ignore human worries about technological revolutions. But that’s a luxury we humans unfortunately cannot afford.