Varun
Ammanagi.

Most of what we read and watch found its way to us through the same mechanism: a popularity filter. When you search for a book on a topic, the results are ranked by copies sold and aggregate ratings. When YouTube's recommendation engine surfaces a video, it's optimizing for watch time across a population. The question worth asking — and rarely asked — is: whose opinion is this, exactly?

The answer is median. Bestseller lists, Amazon rankings, and algorithmic feeds are calibrated to the largest possible audience, which sits near the center of the distribution. That isn't a conspiracy; it's just how optimization works. Platforms maximize engagement, and the mode of the population defines what's engaging at scale. Content pitched too complex loses viewers; content pitched too simple bores them. The result is a vast middle band — readable, accessible, reliably unsatisfying to anyone trying to go deeper.

The same dynamic plays out on YouTube. The videos with hundreds of thousands of views are, by construction, the ones a broad audience found accessible and entertaining. The video with 800 views and a wall of equations in the comments might be the more rigorous treatment. The signal and the noise are inverted from what the interface implies.

None of this makes popular content bad — it makes it a starting point. The problem isn't consuming it; the problem is mistaking it for an endpoint.

This is where AI search, used carefully, breaks the loop. A well-constructed prompt can bypass the popularity filter entirely — asking not for the most-recommended book, but for the current state of expert consensus, the strongest counterarguments, or the papers that practitioners in a field actually cite. The key word is carefully. A casual AI query reproduces the same bias; the model has been trained on the same internet that ranks by popularity. Precision in the prompt is what shifts the output distribution toward the tail.

There's also a subtler trap worth naming. People are generally aware of this dynamic in their areas of competence — in the field you know well, the shallowness of the popular account is obvious. The trap is assuming that awareness transfers. Most of us are operating on surface-level exposure in the majority of our interests, without the expertise to notice. The book that felt comprehensive probably wasn't. The YouTube explainer that clicked was probably Grade 9.

Starting at the median is fine. Staying there is the mistake.

The current state of the equity market doesn't line up with basic sense anymore. A company raises real money only once — IPO or FPO. After that, it's just old shares flipping hands. The guys who gave the company money during the IPO deserve a return — makes sense — they should get paid back through dividends. If a company does well, dividends grow, fair enough. Some premium on the share price for expected growth is logical too. But there's a limit. You can only increase dividends so much, and you can't distribute more than you earn forever.

Yet look around — the premium people are paying today has nothing to do with any realistic dividend or buyback flow. Nobody cares about actual cash returns anymore — it's pure capital gains, fully dependent on someone else paying even more tomorrow. The new buyer isn't paying more for higher dividends — he's just betting the next guy will pay more still.

At this point, it's just a musical chairs game. No real yield, no actual return — just momentum. It can go on for a while if enough people keep showing up with fresh money — cheap debt, low rates, or blind indexing help too. But the math doesn't change — if you're never paid back by the business, the only way out is offloading the bag to the next idiot. When that chain snaps — it's going to be Armageddon.

So right now, the market mostly runs on the idea that you don't need a real return from the business — just a bigger fool to take you out at a higher price. The question is simple — how long can you count on the next fool showing up?

I don't know who will win or lose the AI race, but it's not going to be OpenAI. Claude could, Gemini might, or maybe some new player. But not OpenAI.

Main reason: OpenAI is at a huge disadvantage. Having the first mover advantage is the worst thing that could have happened to it.

There are so many freaking users using their chatbot every day — for free! This, in theory, shouldn't have been a problem because in the LLM world, more users = more training data = better model, right?

But think about it for a second. What sort of data are they getting? Most of it is just redundant. A company can only extract so many insights from its users' engagement; beyond that, it's just unnecessary data.

Whereas with Claude, the average user doesn't even know about Claude. But anyone even slightly enthusiastic about the AI world used it. This not only reduces their burn rate but also gives them better quality data to work with.

Gemini — well, it can go as long as Google wants it to. It is funded by a behemoth sitting on a ton of cash. To give some context, OpenAI's yearly burn of $10B is roughly 10% of Google's annual profits.

We have all heard the phrase "higher the risk, higher the returns" at least once. But is it really true?

When someone says "I am taking a risk," what they mean is the probability of something happening is low or uncertain. Uncertainty comes from not knowing completely. The more you analyze, the higher your probability of being right.

Say I ask you to invest in Devyani Ltd., a newly incorporated company about to start designing its product. You probably wouldn't — you don't know the management, the product, the market fit. Risk is high because you're uncertain.

Now say I ask you to invest in Trent Ltd., co-founded by Warren Buffett and Charlie Munger, with Elon Musk and Tim Cook as Product Heads, Jeff Bezos as CEO, Mukesh Ambani as CFO, and Ratan Tata as MD. You don't know the product either — but you know the people. Your certainty is higher. Risk is lower.

Conclusion: risk comes from not having complete information. The more you know about what you're getting into, the less risk you're actually taking. Our goal shouldn't be to run from risk but to reduce it by doing the work.

Invest your time before you invest your money.

"Risk comes from not knowing what you are doing." — Warren Buffett