The internet was supposed to create a generation of geniuses. For the first time in history, all of human knowledge was available to anyone with an internet connection. Every textbook, every lecture, every research paper, every how to guide. Most of it free. Available to anyone in the world.

This was supposed to be the great equalizer. A kid in rural China could learn the same physics as a university student at MIT. A teenager in Appalachia could master programming without ever stepping foot in a university. If access was the problem, the internet fixed it.

Of course, that was the theory. Does the world today feel to you like it’s full of geniuses?

Here is what actually happened: we got social media feeds, attention economies, and infinite scrolling. The same people that were supposed to self educate to become geniuses, turned out to prefer 15 second videos of strangers dancing.

The Internet’s Broken Promise

Think about the 2010s. Khan Academy. MIT OpenCourseWare. Wikipedia. Stack Overflow. YouTube tutorials on literally everything from quantum mechanics to plumbing. The entire curriculum of a world class education, accessible from a phone.

Some people used it that way. Autodidacts taught themselves to code, to design, to build companies and so on. The internet created real opportunities for the very small percentage of people who had initiative. Everyone else got Facebook.

Futurists and utopians quite often miss this. Access to information is necessary for learning, but it is nowhere near sufficient. Learning requires sustained attention, practice, tolerance for frustration, and the ability to delay gratification. No amount of clicking on Instagram is going to give those to you.

The platforms figured this out very quickly. Engagement became the one metric to optimize. The algorithms learned to serve what people wanted in the moment, which is almost never what would make them smarter, healthier, or more capable. And so the internet became a mirror of human nature, and human nature chose entertainment over education for the vast majority of people.

The Pattern Repeats

Of course this was the same pattern that happened over the entire 20th century. Every major communication breakthrough was the same.

Radio was going to democratize education. In the 1920s, universities launched radio stations expecting millions of eager listeners to tune in for lectures on literature and science. Within a decade, soap operas and variety shows dominated the airwaves. The educational stations went bankrupt.

Television was going to be the great classroom. The FCC reserved channels specifically for educational programming. Newton Minow, the FCC chairman, famously called commercial television a “vast wasteland” in 1961. How did that turn out?

And here we are again. A new medium appears. Visionaries predict it will elevate humanity. The medium gets adopted. It gets used primarily for entertainment and convenience. A small minority uses it for genuine growth. The majority does not. And we have yet another disappointment.

Enter AI Agents, Same Script, New Cast

Now, if you have spent any time on X or LinkedIn you are probably tired of hearing about AI agents. Every job will be automated. Everyone will become a 10x worker. Entire industries will be restructured overnight. We are about to see the greatest productivity revolution in human history, blah, blah, blah.

Of course these are the same exaggerations that they said about the internet . Swap “browser” for “AI agent” and “information” for “automation” and you will not see the difference.

The same predictions. The assumption that access to powerful tools automatically translates into powerful outcomes.

Spoiler alert. It won’t.

Here is what will happen instead. AI agents will become widely available. They will be genuinely capable. They will be able to write code, write documents, analyze data, manage schedules, and automate repetitive tasks. The tools will be real and impressive.

And most people will use them the way they use the internet: for convenience, for shortcuts, for doing the same things slightly faster. They will generate email replies and summarize meetings and create passable first drafts of things they would have written themselves anyway.

A small percentage of people will use AI to do extraordinary things. They will build new products, accelerate research, automate entire workflows, and create leverage that was previously impossible. These people will capture enormous value, no question about that. These are the same kind of people who taught themselves to code from YouTube tutorials in 2012. But the vast majority of people won’t. Mark my words.

The Real Lesson

Technology is an amplifier of intent. The printing press amplified the ambitions of the literate. Radio amplified the reach of the charismatic. The internet amplified the curiosity of the self directed. AI will amplify the productivity of the agentic.

The constraint on human achievement has always been agency. Agency and discipline, not access to tools. Giving everyone access to all of human knowledge did not produce a generation of geniuses. It produced a generation of scrollers and a tiny few polymaths.

AI will follow the same distribution. The people who are already intentional about their work will get dramatically more effective. The people who are coasting will coast with fancier tools. The aggregate productivity numbers will tick up. The revolution will be quieter and will happen slower than most AI pundits are predicting.

So What?

None of this means AI is overhyped in terms of capability. The technology is genuinely transformative. I’m an AI startup founder myself and a true believer in the power of AI to solve human problems.

The mistake is assuming that transformative technology automatically transforms everyone who touches it. It doesn’t now and it never has.

The internet put the Library of Alexandria in every pocket. Most people used it to argue with strangers and watch cat videos. AI will put a brilliant, tireless assistant at everyone’s fingertips. Most people will use it to avoid thinking.

And a few people, the ones who were going to do remarkable things anyway, will use it to do those things faster and at a greater scale. And probably that’s enough.