What happened
The community's first AMA, and the guest was younger than most of the people asking. Shubhram Bhattacharya had left IIT Kanpur in 2021, built two consumer products from scratch, taken one past a million users, and six months before this had quit to start Iterate AI. He is also a state-level 400 metre sprinter and boxes at city level, and most of his answers came back to that.
The idea he kept returning to was controlling the inputs. He splits a goal into steps, sets a target for each step, and then refuses to think about the end. Wanting answers to a set of questions within a week became: find fifty people who could answer them, of whom ten replied, of whom he wanted five on a call. The discipline is from sport, where you train without knowing the result of the competition, and where the practice is the only part you control.
On user interviews he had two rules, both about status. Be honest with the person about what you want out of the conversation and leave no grey area. And go in assuming you know nothing, or that what you know is wrong, because people feel useful when they are helping you and say less when they think you already have the answer. "Please help me out here" gets more than an offer of a trade.
Asked what he looks for when hiring, he named two things and neither was a qualification: an extreme hunger to learn, and enough shamelessness to admit what you do not know. Pressed on how somebody without a product background is meant to demonstrate the first of those in an interview, he gave the worked example. Before interviewing at Stable Money he learned the company, found its competitors, spoke to product managers at those competitors to guess what their next quarter looked like, and estimated Stable Money's revenue, which meant sizing the market for buying fixed deposits online and finding out what commission a bank pays a platform that sells one. He also spoke to a junior person inside the company about what was bothering them that month. He was roughly right rather than right, and said that was not the point.
He was asked whether AI makes product management a bad bet and did not think so: fields get augmented and renamed rather than deleted. The AI app developer he is building is meant to take the work app developers dislike, the copy changes and the SDK integrations, so they can spend the time on the engineering that made them want the job. On startups dying when a large company announces something, he said most of them pivot rather than die, and that the vision is what you hold while the approach is one of a hundred routes to it.
The most revealing answer was about why he works this way at all. He said he does not like to relax, gets bored within a day or two of decompressing, and treats rest as preparation for the next fight. He also talked about his father, who had an electrical diploma and spent a career at the state electricity board, and who ended up the person the city called about anything to do with electricity supply. Not fancy and not lucrative, and the respect was real.