Why AI won't make you 10x smarter (and I keep buying tools I don't need)
I have a problem. I keep buying tools I don't need, convinced this time will be different. A new note-taking app. Another todo list system. A fancy AI wrapper that promises to "supercharge my workflow." I download it, configure it, use it for two days, and then I'm back to exactly where I started, minus forty dollars and maybe a little sadder.
The AI version of this looks like: "If only I could wire up the right prompts and the right tools, I could have an agent that boosts my productivity 10x, or fixes my problems, or makes me more knowledgeable." There is a line that goes around on Twitter, and I have never managed to trace it back to whoever said it first, that puts it well: your laptop has a hundred-million-dollar startup in it, you just have to figure out the right sequence of words to get it out.
This idea is not new. The early computing pioneers had the same ambition. Engelbart literally called his lab the Augmentation Research Center. The dream of using computers to amplify human intellect is almost as old as computers themselves. And yet here we are, still waiting.
Why? I think there are two reasons. First, most of us don't actually have a concrete need that AI would solve. Second, even when we do, the real bottlenecks are inside our heads, where no amount of clever scaffolding can reach.
The "I Want an AI That Writes Flashcards" Problem
Let me be honest about something. I've heard dozens of people, online and in person, tell me some version of: "I want an app that uses AI to write flashcards."
How many of those people do you think have ever written a flashcard by hand? How many of them use Anki every day? Have ever used Anki at all?
The people who want an AI that writes flashcards for them don't use flashcards. They have no reason to. And the people who DO use flashcards wouldn't benefit much from AI writing them, because the act of writing the card is half the learning.
This pattern repeats endlessly. AI tutors. AI executive assistants. AI that curates your "digital garden." AI that writes your notes for you. The template is always the same: if you could just find the right incantation, your life would transform.
I don't think that's how it works.
Take the AI tutor fantasy. Imagine you had the ghost of John von Neumann living in your laptop. What would you have him teach you? Let's be real. You'd go through chapter one of some math topic you're vaguely curious about, get busy, forget about it, and never open it again. And that might actually be the rational choice. Most people are not autodidacts. Not because they're lazy. They just have no material reason to learn a specific topic. Their job doesn't require it. And the problem with learning for the sake of learning is opportunity cost: there's no reason to pick one thing over another, so the rational move is to do nothing and wait for something that actually grabs you to appear.
Would an AI executive assistant actually boost your productivity? What would it do, other than tell you to do the things you already know you need to do? The way to deflate these ideas is to get concrete. Take a day at work and ask: what exact actions could an AI looking over my shoulder have taken, that would have made a real difference? If the answer is "not much," you don't have a tool problem. You have a needle problem. There's no needle to move.
The Second Brain That Doesn't Need to Exist
I say this with love because I am a recovering tools-for-thought person myself. I hoard PDFs. I have folders full of articles I will never read. I have tried more note-taking systems than I care to admit.
Most of us, unless we are students, don't really have a reason to take notes. If you're a student, you take notes from the textbook. If you keep a journal, that's occasionally useful. If you stopped, probably not much would change.
The notetaking people I'm about to describe are never doing hard work. They're not a researcher at the frontier of their field, building deep expertise over time. They're not a historian who has to read millions of words across thousands of sources to reconstruct someone's life. It's always some blogger. Their "digital garden" is about how to keep a digital garden. It's very solipsistic. There is no output, no deliverables. The deliverable is a screenshot of an Obsidian graph posted on Twitter. It looks like an incomprehensible ball of twine. Everyone in the replies says "nice graph." Nothing has been accomplished.
So what difference is the AI going to make? Write my notes? About what? Read articles and summarize them and add them to the garden? For what purpose? Find connections between my ideas? What ideas? An AI pulling together an unfinished list of bullet points for a draft of an essay about some inane thing, plus a bunch of PDFs I haven't read, does not produce anything of value. It produces another project I'm not going to do.
I used to be this guy. I was deep in the tools-for-thought ecosystem. I spent more time organizing my notes than actually thinking. But let's be honest with ourselves. Sometimes the tool doesn't move the needle because there's no needle to move. The "needle" is not a concrete, realizable need. It's a vague, aspirational idea about who we want to be.
The Bottleneck Is Inside My Head
Let's say you do have serious work. You have real problems to solve. The AI scaffolding is in place. Why aren't you suddenly 10x more effective?
Because the real limits are inside your head. You can build all the scaffolding you want around a human, but the human stays the same.
I have ADHD. I've built systems. Todo lists, calendars, timers, a million little ways to trick myself into working. This gets me from "completely non-functional" to "kind of functional." And then it saturates. No amount of pomodoros can do what a little molecule diffusing through my brain tissue can do. All the productivity systems in the world are as nothing compared to actual medication targeting the real bottleneck.
Is an AI agent going to fix that? I doubt it.
What about intelligence? Can AI augment human intelligence? What does that actually look like? AI agents only became useful when the models crossed a certain capability threshold. You can't put GPT-2 in a fancy harness and get GPT-5 results out of it. Can you put a human in an AI scaffold and give them plus thirty effective IQ points? I don't think so. Not unless the AI is doing all the thinking, in which case what use is the human? The limiting factor in the human-AI partnership is the human. Intelligence is fixed until we figure out some specific biotechnology.
And then there's knowledge.
The Invisible Ceiling
Even very educated people underestimate the role of knowledge. There's this attitude that you can just Google everything just-in-time as it comes up. I honestly don't understand how anyone can think that.
Everything you do, every action and idle thought, draws on knowledge you don't even notice is there. Here's my favorite example. Claude Shannon laid most of the foundation of digital computing because he remembered this then-obscure branch of mathematical logic called Boolean algebra and showed, in his 1937 master's thesis, that it mapped straight onto switching circuits. A trillion-dollar industry, conjured out of some old books he read.
The reason knowledge is still a bottleneck in the AI era is not "if you don't have the knowledge, you can't write the prompt." If you don't have the knowledge, you don't understand the question. You don't know why it matters. You can't judge the answers. You won't even think to ask. You are in a completely different continent from "writing the prompt."
AI cannot inject knowledge into your brain. Long-term memory is private and internal. It can maybe help you acquire new knowledge, if you use it carefully. But it cannot give you the thing that makes Shannon's insight possible: a deep well of concepts and connections that let you see what nobody else sees.
So What Does This Mean?
The three big bottlenecks are executive function, intelligence, and knowledge. All three live inside your head. AI cannot touch them until we have biotech that changes the hardware.
The returns to education are actually higher now. The popular narrative says human capital is becoming worthless. I think the opposite is true. People who already have knowledge and discipline stand to gain more from AI than anyone else. Because they already have something to augment.
For the rest of us, well. We have to do the boring work. Build knowledge. Fix our executive function. Develop discipline. The AI is a tool, not a shortcut. A powerful tool, maybe the most powerful one we've ever built. But a tool still needs someone who knows what to build.
I keep buying tools I don't need. But at least now I know why.