The morning briefing everyone wants to build

Summary: Readers didn't write in about the $80 billion AI bet. They wrote in about something smaller and harder. One reader put it precisely: 'I don't know what to look for in a factsheet. So I won't know what to ask the tool either.' That gap is the real conversation.

The responses to Dhirendra Kumar’s latest Editor’s Note, The AI revolution worth having came quickly and from a wide range of people. Most readers had little interest in relitigating the $80 billion debate. What struck almost everyone, it turned out, was the smaller thing: the morning briefing, the audience of one, the idea that a person of any age could build something genuinely useful for themselves. That is what they wrote in about.

There was a question almost everyone asked first.

It arrived in dozens of variations, from readers across India and beyond, phrased with varying degrees of urgency. Sujatha Rao, a disciplined investor of two decades, wanted to know which specific tool was being referred to. Sunil Kothare asked whether Copilot, Claude, Gemini or Perplexity was the right place to start. Krutik Sanghavi, 32 years old and investing for 14 years, was the most direct of all: "How do I make one for myself?"

The question is understandable. It is also, in one sense, the wrong question.

What the note described was not a product you download and immediately use. It was something built over weeks, through patient instruction, trained on what to watch and what to ignore, shaped by the specific priorities of one particular investor. Pradeep Kumar Nair came closest to understanding what was actually on offer. "You have given me direction," he wrote. "Something I will definitely do." The direction, not the tool name, was the point.

But before direction comes a harder question, one that most readers approached carefully and only a few addressed head-on.

Before you can ask, you have to know what to ask

Abhijit Chakravorty put it plainly, and it took some honesty to do so.

"A tool can summarise a fund factsheet in a minute," he quoted back. Then came the real observation: "I just don't know what to look for in a factsheet. In that case, I won't know what queries should be created for the tool to give me the right kind of output which is of relevance to me."

He called himself "one such ignorant investor," which is probably too harsh a verdict. Anyone who can identify the problem that precisely is already some distance ahead. But the point itself is important. These tools are genuinely powerful. They are also only as good as the questions you bring to them. A plain-English summary of a factsheet tells you what the document says. Whether it helps you decide anything depends on what you already know to look for.

Bharg Joshi, who described himself as both an IT student and a market participant for five years each, made a related point with more directness. The people who will use AI well in financial markets, he argued, are the ones who already understand markets. "Even after a few years when AI models would not be a new thing, there will be people who won't make efficient use of these models because they themselves have never educated themselves on markets. So they have nothing to teach or train the model."

It is worth sitting with that. The note described weeks of patient teaching, of telling the tool what to watch and how to speak. That knowledge had to come from somewhere. The tool did not supply it.

Jaspal Kaur Virk, who began investing around the time of the 2008 crisis, put the danger most vividly: "AI is so powerful that if the user is not sure of what it needs, then the AI takes over and the user becomes a tool for AI to show its strength. We all need to learn how to tame the beast."

Ankit Nanda, already a regular user, was candid about what the other side of that trade looks like. The tools have amplified what he can accomplish, he said, "at the risk of losing my own voice. But it's a trade I am willing to make, many times over." That is a reasonable position. It just requires knowing, going in, what you are trading.

The bigger debate some readers couldn't step away from

The note made a deliberate choice to stay out of the grand argument: who wins, who loses and whether $80 billion is conviction or madness. Not every reader was willing to follow.

Arka Mandal, who described more than a month spent tracking recent AI developments closely, wrote the most detailed response by some distance. He raised concerns about what he described as regression in successive model releases and drew attention to a phenomenon now being discussed in the American technology press under the name "tokenmaxxing," where engineers at companies reportedly run AI processes not because the work requires it but to justify their organisations' AI investments. He also raised a structural point about Alphabet's position that is difficult to argue with: the company's core business depends on the friction of searching and clicking through links. Large language models are designed to remove that friction. "If it does not develop its own large language models," he wrote, "then other LLMs will eat away its search share. If it does develop them, then they will end up eating away its own search share." His personal conclusion, characteristically, was pragmatic. He had moved his entire workflow to a local open-source model running on his phone at no cost, having decided he could not justify a premium subscription for 20 features when he only needed one or two. It is, in miniature, the same argument the note made: the personal use case matters more than the global bet.

Ajay Tiwari preferred the longer view. Electricity and the internal combustion engine both arrived in the 1870s, he noted, and both took decades to reshape industries in ways nobody had predicted at the time. "Timing or predicting its impacts should be done after a few years of development." His point was about patience rather than pessimism. AT&T jumped in and failed, he acknowledged. But the technology itself stayed.

Pankaj Dubey noticed something counterintuitive. The technology workers most anxious about AI are, by his observation, the first to use it most aggressively. "Survival of the fittest instinct is paying out in full swing. These same individuals are going to be the savviest people to leverage AI because they are the subject matter experts in their area." Productivity gains, he said, may shrink some roles and expand others. Nobody knows the net. But he also flagged something worth noting: the visible hurry with which some AI companies appear to be pursuing IPOs struck him as a sign that the people closest to this technology may harbour their own doubts about how long the current moment lasts.

The small revolution, already underway

While the larger argument was being relitigated, a number of readers wrote in with something simpler. They were already doing it.

Rory Hancock, writing from London, mentioned he had uploaded his entire transaction history to Claude and asked it to compare his active investing record over a couple of years against a passive approach. "It has been insightful to say the least," he wrote, with the slight understatement of someone who may have discovered something uncomfortable.

Suresh Parikh, 80 years old and also in London, enrolled in weekend AI classes organised on Indian Standard Time, got up at 4 am to attend, sat through eight hours of lectures alongside IT professionals looking for job opportunities and learned how to build a video advertisement for an Xbox his 14-year-old granddaughter owns. He has no interest in Xbox. He nonetheless derived what he called "perverse pleasure in learning how I could do things in which I am totally uninterested." There is something about that sentence that captures the spirit of the whole exercise rather well.

Muralidhara Reddy, 60, described using AI for stock evaluation, kitchen queries and health questions with equal ease. Then he added a line that brought the investment argument back to its simplest form: "Who wins the bet is immaterial. It is better to stay within the circle of competence." He does not feel equipped to analyse AI companies or IT services stocks. He invests in businesses he understands. "That keeps things simple, plain and clean."

Anand Kumar recalled learning MS-DOS commands at a computer institute near Banaras Hindu University, the institution's full name—Technical and Social Research Institute—apparently quite imposing for a place that issued three-month certificates. "Whatever we learn is the best return guaranteed," he wrote. The certificate sits alongside his university degrees. Both are close to him.

Jayesh Nachnani closed the loop most neatly. He had decided to follow the editor's example and spend a little time each day going deeper on equities he cared about. "Control the controllables," he wrote, invoking the Stoics. It is not a bad principle for markets or for this particular revolution.

The responses, taken together, do not resolve the big debate. Nobody expected them to. What they do is sketch a portrait of a large number of ordinary investors who are watching the AI arms race from a comfortable distance, who have no strong view on who will win it and who are quietly getting on with the personal version of it anyway.

Some are uploading transaction histories in London. Some are running local models on their phones. Some have paused to ask whether they know enough yet to ask the right questions. That pause is not failure. It may be exactly the first step the note was pointing toward.

The $80 billion question will get answered by markets and by time. The smaller one is more immediately available. It starts with figuring out what you actually want to know. The tool, as several readers discovered, turns out to be the easy part.

Credits

Sujatha Rao, Sunil Kothare, Krutik Sanghavi, Pradeep Kumar Nair Palliyill, Abhijit Chakravorty, Bharg Joshi, Jaspal Kaur Virk, Ankit Nanda, Arka Mandal, Ajay Tiwari, Pankaj Dubey, Rory Hancock, Suresh Parikh, Muralidhara Reddy, Anand Kumar, Jayesh Nachnani

Also read: Helping hands, hidden barriers

This article was originally published on June 10, 2026.