In 1934, the poet T. S. Eliot wrote two questions that read as if they were written for the AI age. They appear in his Choruses from “The Rock”:
“Where is the wisdom we have lost in knowledge? Where is the knowledge we have lost in information?”
Ninety years later, those lines describe our daily working life with uncomfortable precision. We have more information available to us than any generation in history. A large language model will hand you an answer to almost any factual question in a few seconds, at no cost. And yet most of us do not feel ninety years wiser. The questions Eliot asked point to a problem that more information cannot solve, because the problem was never a shortage of information in the first place.
Three different things, often confused
It helps to separate three words we tend to use as if they meant the same thing. Information is raw material: facts, values, documentation, search results. Knowledge is information that you have organised, connected, and made your own, so that you can reason with it. Wisdom is knowing what to do with that knowledge in a real situation, including what to leave out and when to act.
More information has never automatically become more knowledge, and knowledge has never automatically become wisdom.
Each step up requires work that the previous step does not provide for free. Reading a manual page is not the same as understanding the subsystem it describes. Understanding the subsystem is not the same as knowing whether to use it in the design in front of you. Eliot’s two questions are really one observation: the conversions between these levels are where the value is created, and those conversions take attention and effort. They do not arrive in the download.
Why this matters more now, not less
It is tempting to think that AI tools have made Eliot’s worry obsolete. If the information is always one prompt away, why hold any of it in your head? The answer is that the tools have changed which skill is scarce, and therefore which skill is valuable.
When answers are cheap, the ability to turn information into knowledge, and knowledge into judgment, is what stays scarce.
An AI assistant is very good at the first level. It retrieves and reformats information at remarkable speed. What it cannot do for you is the conversion work. It cannot decide that one of the facts it returned is the one that actually matters for your bug. It cannot carry the weight of a system you have built and maintained for two years. It cannot tell you that the textbook-correct answer is wrong for your hardware. Those judgments come from knowledge you have internalised and wisdom you have earned, and there is no shortcut that skips the earning.
For engineers and students this is good news, not bad news. The part of the work that is hardest to automate is exactly the part that was always the most valuable: the human conversion of information into understanding and understanding into sound decisions.
How to do the conversion work on purpose
If the value lives in the conversions, then your study and work habits should protect those conversions rather than skip them. A few concrete practices:
- Generate before you look up. When you hit a problem, attempt your own answer first, even a rough one. Then use the AI tool to check and extend it. You keep the struggle that builds knowledge, and you still get the speed.
- Rebuild from memory. After reading documentation or an AI explanation, close it and reconstruct the idea in your own words or your own code. If you cannot, you have information, not yet knowledge.
- Ask why the answer is right. An answer that works is information. Knowing why it works, and where it would stop working, is knowledge. Spend the extra few minutes on the why.
- Connect new facts to what you already know. Knowledge is information with connections. Deliberately link a new concept to one you already understand instead of storing it as an isolated fact.
- Sit with the hard problem before reaching for the tool. The discomfort of not knowing is where judgment is built. Give it ten minutes of honest thought first.
None of this means rejecting AI tools. It means using them as instruments for the conversion work instead of as a way to avoid it. Outsource the typing and the retrieval. Keep the thinking.
The download gives you information. Only attention and effort turn it into something you can actually think with.
This is the same conviction behind how we teach at TECH VEDA: we build understanding you can reason from, not facts you can only repeat. Eliot’s questions are ninety years old, but they are the right questions to keep asking yourself every working day. When the information is free, your task is no longer to gather more of it. Your task is to do the quiet, deliberate work that turns it into knowledge, and your knowledge into judgment that is genuinely yours.
— Raghu Bharadwaj
