The Autodidact;
In Search of the
Lost Chord

By looking at me, no one ever asked the question at a dinner party or after a meeting at a federal agency in downtown DC: How did you ever listen to that kind of music growing up in India? Perhaps because the people who attend dinner parties or work at federal agencies have never needed to ask: What do you do when you can't find the answer? Get moodily blue?

Not the answer to a trivia question. Not the answer to a crossword clue or a tax form or a GPS reroute. I mean the answer to the kind of question that life hands you without warning — the question of how to survive a new country, or an empty bank account, or a system built for people who already know how it works. The kind of question where the blue links never quite load, and when they finally do, you click and land on a page you don't need.

I have been searching my entire life. For what? Well, everything really, though not metaphorically. Literally, urgently, with the specific desperation of someone who has learned that the cost of a wrong answer is not inconvenience but consequence. And when someone recently told me AI has ended the age of search, I begged to ask: for whom?

The truth, which nobody in Silicon Valley says out loud, is that not all searches were created equal.

The Word I Didn't Know I Was

The internet arrived, as it did for most, but it was a boon most notably for autodidacts like me — people at the center of their own learning, exploring different approaches to knowledge, driven mostly by obsessive curiosity and a questionable ability to resist clicking on every shiny link. Yes, I am still living with chronic pain, and yet somehow I have accumulated some extraordinary entrepreneurial adventures under my belt — of the stranger-than-fiction variety — some involving the usual setbacks, and just the right amount of success to say it was all worth it, all through the pain and the relentless march of technology. I've tried to make sense of it all, often feeling like I was decoding a map that kept redrawing itself.

That's not all. I've spent my life chasing answers to questions that most people only ask after their third cup of coffee or a bad day at work. Why are we here? What's the universe up to? Why does time feel so slippery? Is consciousness just a fancy word for being confused? And, more importantly, will artificial intelligence eventually outsmart us, or just help us find our lost keys?

I recall that asking questions like these made adults sigh, and other kids edge away at recess.

I am going to admit something I find simultaneously embarrassing and entirely consistent with who I am: I did not know the word autodidact until relatively recently. This is, of course, the most autodidactic thing possible — to be a self-taught person who had to teach himself the word for self-taught people, arriving at it decades into the practice, the way a natural swimmer might one day overhear someone describe the butterfly stroke and think, oh, that's what I've been doing.

An autodidact is a person who is self-taught — someone at the center of their own learning, driven by curiosity rather than curriculum, navigating knowledge without institutional scaffolding, building frameworks out of available materials and stubborn questions. The moment I encountered the word, I felt the particular relief of a person who has been describing a colour in circumlocutions for years — you know, that blue that isn't quite blue, that has some green in it but not quite — and is then simply handed the word teal. I had been teal my entire life without knowing it.

What I did not realize, in the moment of relief, was that recognizing the word for what you are does not retroactively organize your trajectory into something that looks like a plan. If you were to draw a map of my intellectual and professional life, it would resemble less a journey and more a Roomba navigating an unfamiliar kitchen: confident, purposeful-looking, occasionally stuck behind a chair leg, sometimes returning to exactly where it started as if verifying something, and moving at all times with tremendous conviction in directions that turn out not to be the direction.

The Library That Never Closed

Suddenly the library of the world was free, accessible with simple mouse clicks, and you could click away into the night. Suddenly the reference desk never closed; you could always find it open at 4 AM. Suddenly a name like mine, which had the uncanny habit of announcing itself before I could — was just a cursor in a search box, indistinguishable from any other. I was an addicted autodidact with a mouse under my right paw, an index finger clicking away to glory.

I want to tell you that I used this power wisely and efficiently. I want to tell you that I typed clean, precise queries, evaluated sources with scholarly dispassion, and arrived at answers with the calm of a man who had always had access to answers.

It didn't always go that way, to be honest.

I once spent hours trying to determine if a work contract I received was legal. I searched through odd message boards and concluded it probably wasn't. I then spent additional hours chatting on a forum with someone who claimed to have the same situation. Without solid evidence, I decided the advice might be applicable and acted on it — with a margin of error that, in hindsight, I am genuinely surprised did not ruin me.

This is not a complaint. This is a data point.

Because autodidacts generally don't learn in any classroom; they only ever learn through trial and error. Fall and get up.

What I have come to believe is truly one of the most underappreciated gifts of a life that did not go smoothly until it finally does: adversity teaches you, amongst other things, to search for meaning. Not to retrieve results. Not to browse aimlessly. But to search — with intention, mostly with skepticism, because I had the earned understanding of someone who knew internally that the first answer would almost never be the right answer for me, and that the quality of what I find depends almost entirely upon the quality of what I ask. That clearly demanded some formal education, and thankfully, I was fortunate to have one. That was the defining moment of my understanding why my parents always stressed education. That I felt thankful hardly captured the emotion that hit me. It has stayed with me with unflinching intensity ever since.

The Archaeological Record

I am writing this book — History's Future — and talking through a website I built myself, not because I couldn't afford to hire someone to build it for me. It is because I learned over time that I was a born autodidact, a self-learner, driven to earn my very own confidence from my inner conviction that nothing was ever that hard to accomplish if you just tried.

Reader, I tell you honestly, it was very, very hard. Nothing is easy; I can underline that for you.

Reader, it was that hard.

There is a server in the Amazon cloud somewhere in a datacenter in Virginia that contains what I can only describe as the archaeological record of my autodidactic technical trajectory. It contains stratified layers of many wrong decisions, each sealed beneath some good idea that came next before I could understand why it preceded a wrong one.

At one point, I had forty-four subscribers to my newsletter. I was briefly delighted by this. I told people I had forty-four subscribers. It turned out forty-three of them were spam bots. I am not certain the forty-fourth was human. I purged the bots and was left with what is technically called a "small but engaged audience" and what is practically called talking to yourself with a mailing list. That one name was mine on that list.

Then, of course, due to the mistakes I made configuring my AWS server, my email system stopped working. Not immediately — that would have been merciful. It failed silently for over twenty days while I was dreaming, wearing a happy but delusional smile, believing my words were reaching people. They were reaching no one. Zilch. My server was logging errors at 2 AM while I slept through them all. When I finally discovered the outage, I sat with a peculiar stillness — that frown of a person who, despite having spent decades learning that systems fail, that entropy always wins, was somehow stupefied by it all over again.

I rebuilt the email system. I added monitoring. I wrote documentation detailed enough to feed ChatGPT or a future Claude so it could understand what had happened, what it had done, and why. I was done spending time building and rebuilding what AI could now handle. After all, this entire writing project is about AI.

Perhaps in the doing of all of it, I asked better questions than I had asked before — not because I became smarter than I used to be, but because I had paid, in time and embarrassment, for the specific knowledge of where my previous questions had gone wrong.

Such is the curriculum adversity offers, and believe me when I tell you that it is not available in any other format.

What AI Has Amplified

The great intellectual promise of AI is synthesis. Where the search engine retrieved, the model reasons. Where the index ranked pages, the AI distills and responds. This is genuine. I use these tools every day. I have watched them compress what would have been a week of research into an afternoon, surface connections I would not have made in a month of reading, and explain technical concepts I had been too embarrassed to admit I didn't understand.

But there is something troubling about all of this. I suspect this is not unique to me: the people who get the most from these systems are the ones who have already learned to search badly.

By which I mean: they have searched badly enough, and paid for it enough, to understand that the framing of a question is not a preliminary step to thinking — it is the thinking. That "how do I fix this error" is a worse question than "what does this error tell me about my assumptions." That "what should I do" is entirely much worse than "what are the three ways this could go wrong." That the system, any system, will reflect back the quality of your inquiry with the kind of fidelity that, frankly, will make you feel foolishly uncomfortable.

The person who grew up with a reference librarian, a well-funded school library, and parents who knew which questions to ask — that person had help calibrating their inquiry from the beginning. For the person who had a different kind of education: the hard curriculum of learning, through setbacks, through falling and getting up, the search is only as good as the searcher.

What AI has done is not eliminate this distinction. It has amplified it.

The well-formed question now gets an extraordinary answer. The vague question gets a fluent but useless one. The person who has never learned to interrogate their own assumptions gets a very confident response to the wrong question. The person who spent years searching badly, and knows it, brings to the AI something that the AI cannot generate for itself: the scar tissue of a harder search.

The Bottleneck That Always Mattered

History has always rewarded the precision of questions over the abundance of answers. The printing press did not make everyone a scholar; it made the already-curious more inquisitive and formidable in their quest for knowledge, while the incurious simply became comfortable in their ignorance. The internet did not flatten knowledge; it gave people with existing knowledge frameworks the raw material to build further, while everyone else got a great deal of content to scroll through. AI will not be different. What will be different is the speed at which the gap between the well-formed question and the poorly-formed one becomes visible.

I don't like to call it prompt engineering. That sounds too contrived. I call it engineering to prompt.

There is an old idea, usually attributed to circumstances more dramatic than mine but applicable to all of them, that difficulty is a form of instruction. That the person who has had to fight for an answer — who has been turned away from the door and had to find another door, who has been given the wrong map and had to triangulate from landmarks, who has built the wrong email system and had to understand why — carries something that the person who was given the right answer cannot easily acquire.

I am not going to romanticize this. Difficulty is also just difficulty. I would have preferred a functioning newsletter from the beginning. I would have preferred not to have spent hours reading message boards that probably didn't apply to me. The sentimental argument that hardship is secretly a gift is the argument of someone who got through it, and I am aware that not everyone does.

But I will say this: somewhere in that archive of wrong turns — the forty-three bots, the silent email outage, the server in Virginia that still contains at least three copies of a nav bar I no longer use — there is an education in the structure of problems that no tutorial provides. I know what my questions cost. I know what a vague question costs. I know the specific texture of an answer that sounds right but isn't, because I have acted on it and learned the difference.

This is what a tough life searches for, without knowing it is searching: the capacity to ask, precisely, what needs to be known.

The Chord That Never Resolves

The end of the search, as I argued in these pages, is the end of retrieval as the primary cognitive act. The synthesis happens first now. The model reasons before the human frames the question fully. What remains — what has always been the deeper act — is the judgment about which question to ask.

For those of us who learned to search the hard way, this is not a loss. It is, if anything, a kind of vindication — the belated discovery that the thing adversity was teaching us all along was not how to find answers, but how to deserve them.

For the first time in history, the quality of a question is the only bottleneck that matters. Which means, for the first time in history, the people who spent their lives learning to ask better ones are holding an asset that compounds.

Search has a long moody chord; it never ends. It always seems to flip around, never gets lost, but starts anew.

I won't ever be done searching.