Seven hundred years ago, if you wanted a book, you paid a person to make it for you, one line at a time. A skilled scribe would spend months bent over a single volume, and for anything elaborate it could take more than a year. The price reflected that, because you were essentially buying a year of a craftsman's life. Knowledge existed, but it was rationed by the sheer human labor required to copy it.
Then Gutenberg came along, and the cost of knowledge took its first real step function down. Once a page could be set in type and pressed again and again, the price of books fell by around 2.5% a year for more than a century, and every time a new print shop opened in a city the local price dropped by roughly a quarter. A book went from costing several weeks of wages to less than a single day. That is not a gentle improvement, that is a step function change, and it is roughly the moment that ideas stopped belonging only to the church and the very rich.
The next step function change was the public library. Andrew Carnegie funded over 2500 of them around the world, more than 1600 in the United States alone, because he believed the best gift you could give a community was free access to knowledge. And here is the part that I think gets overlooked. The library dropped the money cost of knowledge close to zero, but it did not make knowledge free, because now the cost was your time. You had to physically go there, work out the rough category your answer lived in. Then the Dewey Decimal system would point you at the right shelf for business or music or art, and then you pulled book after book and read until you found the paragraph you actually needed.
Google was the next step function change, and I do not think it did the same thing as the library, because what Google really solved was indexing. The library could point you at a shelf, but a shelf is a blunt instrument, since it groups things by broad subject and then leaves you to hunt. Google indexed the actual contents of the page, so for the first time you could ask for the specific thing you wanted rather than the category it might be filed under. That was a big leap, and while the cost of Google was basically nothing, you still paid in time. Trawling through ten blue links and skimming three articles to find the one sentence that answered your question. Better indexing, faster than a library, but still an indexing problem.
If you look at this you can effectively say two things have been falling together. Gutenberg killed the cost of copying. Libraries pushed money close to zero but left you paying in time. Google attacked the time cost by indexing the world's pages so you could find the right one faster. Step by step, on both fronts at once, money and time, the cost of reaching a piece of knowledge has been grinding down toward nothing.
AI is the step function that finishes the job, and it does it by being the best indexer we have ever built. When a model chunks a document into pieces and turns each one into a vector, what it is really doing is indexing meaning rather than keywords, so it can retrieve the precise idea you need even when you never used the right word for it. And even the cost of this continues to fall. The cost of running a model as capable as the one behind the first ChatGPT fell from about $20 per million words of output in late 2022 to about $0.07 two years later, which is a roughly two-hundred-and-eighty-fold drop. So for the first time both the cost of time and money have hit the floor together. The entire knowledge of the world is now at our fingertips in seconds, because something does the going, the getting, and the indexing for you, for fractions of a cent.
Which changes the equation entirely.
For seven hundred years the problem was always some version of the same thing, namely, what it would cost you, in money or in time, to go and find the information. That problem is essentially solved. So if reaching knowledge is no longer the constraint, what is actually left that is scarce and worth anything?
There is a simple idea in economics that tells you where the value goes when something becomes free. It does not disappear. It moves to whatever sits next to that free thing and is still scarce. Think about music. You can now stream almost every song ever recorded for the price of a sandwich a month, so the recording itself is worth very little. The value did not vanish, it moved to the one part that stayed scarce, the live show, which is why musicians now make their living from touring rather than from records. Strategists have a name for this, commoditizing your complement.
Knowledge is now that free thing. So the question worth asking is what sits next to it and stays scarce, because that is where the value has quietly moved. I think there are two such things, and they map almost perfectly onto how a career or a company actually succeeds or fails.
The first is the question.
An index is only ever as good as the thing you ask it to find, and that has quietly become the whole game. If you cannot frame the right question, the most powerful indexing system ever built cannot help you, because it does not know what you are actually trying to find. The quality of what you get out is now capped entirely by the quality of what you ask. We have spent a generation rewarding people for having the answers, and we are walking into one that rewards people for having the right question. Those are not the same skill, and plenty of people who are excellent at the first are surprisingly poor at the second.
The second scarce thing is action.
This is the one I think people most want to ignore, because knowing something has never been the same as doing it. Almost every smoker knows that smoking will hurt them. The knowledge is total, the results are in, the warning is printed on the box in letters you cannot miss, and yet the cigarette still gets lit. The knowledge changed nothing because it never crossed the gap into action. Management researchers gave this a name, the knowing-doing gap, and they found that companies pour billions into knowing how to improve while changing very little about what they actually do (I have made this mistake). When knowledge was scarce and expensive, simply having it was a real advantage. Now that anyone can summon it instantly and for nothing, having it is worth almost zero, and the entire advantage sits on the far side of that gap, with the people who actually act on what they learn.
So here is what I would do with all of this, whether you are building a company or building a career. Stop competing on what you know, because that race is already lost to a model that knows more than you and costs less than your coffee. Get deliberately good at asking sharper questions, since that is now the binding constraint on everything you can pull out of these tools. And then close the gap, relentlessly, between the moment you learn something and the moment you act on it, because that gap is where almost all the remaining value now lives. Knowledge is going to zero. The question and the action are the only things that aren't.