Krishna Agarwal

1. This isn't fake. 2. Credentials are stored as hashes. It should be literally, with no exaggeration, impossible for a vendor to know your credentials while uppercase UNLESS they weren't storing passwords as hashes. What the fuck is HSBC India doing?

Worth asking btw if code fully written with AI has copyright protection? Anthropic is big in saying most of Claude Code is generated by AI. Not 100%, but close, and the last few months pretty much all. Another interesting copyright question.

India ran the most important cardiovascular study of the 20th century by accident, and then immediately forgot about it. In 1967, Dr. S.L. Malhotra published a study in the British Heart Journal examining heart disease rates among 1.5 million Indian railway employees. The population was extraordinarily useful for research purposes: same employer, same healthcare access, comparable income and working conditions, spread across the entire country. The only meaningful variable was geography. Which meant diet. North Indian railway workers: Punjab, Rajasthan, UP, ate a diet built around ghee and dairy fat. They consumed up to 19 times more fat than their southern counterparts. The fat was primarily saturated: clarified butter, milk fat, the short-chain saturated fatty acids that Ancel Keys had recently been telling the Western world were arterial death. South Indian railway workers ate a diet based on rice, sambar, and seed oils: groundnut oil and sesame oil, primarily. They ate considerably less fat overall. By the standards of dietary advice being formulated in the 1960s, they should have been the healthy ones. Heart disease mortality in South India: 135 per 100,000. Heart disease mortality in North India: 20 per 100,000. Seven times higher in the population eating seed oils. Among railway sweepers specifically, the lowest-paid, most physically active workers, the gap was even wider. Heart disease was fifteen times more common in the South Indian sweeper population than in the North Indian sweeper population. Malhotra controlled for everything he could reach: smoking, where Northerners actually smoked more. Activity levels, where the relationship was inconsistent. Socioeconomic status, where executives died more often than sweepers regardless of region. He found no variable that explained the gap except the type of fat in the diet. He published the data. In a peer-reviewed journal. In 1967. The study was cited periodically, acknowledged as methodologically interesting, and then set aside. The decade in which Malhotra published was the decade in which Ancel Keys's fat hypothesis was being converted into policy. The American Heart Association was issuing guidance recommending polyunsaturated vegetable oils as replacements for saturated animal fats. The food industry was producing seed oils at industrial scale. The infrastructure of seed oil promotion was being built, expensively and with great institutional momentum. A study showing that populations eating animal fat had a fraction of the heart disease of populations eating seed oils was not, in that context, a study that anyone particularly wanted to follow up. Nobody followed up. Almost sixty years later, the finding stands unrefuted in the literature. It is not in the dietary guidelines.

Mark Zuckerberg just described the minimum viable business for the next decade. A fourth item made the checklist. Zuckerberg: “Every business, just like they have a website, and a phone number, and an email address, is also going to have an AI.” Website. Phone number. Email address. AI agent. That is not a prediction. That is a new baseline. Twenty years ago, not having a website was a choice. Then it stopped being one. Nobody scheduled that transition. The same filter is back. Running faster this time. A business without an AI agent handling sales, support, and customer interaction will not look outdated. It will look abandoned. Its competitor’s agent responds in two seconds. Knows every customer by name. And while it’s handling yours, it’s handling ten thousand others. You do not outwork that. You do not outspend it. You just lose to it. But Zuckerberg went somewhere most tech CEOs refuse to go. He picked a side in the debate most CEOs avoid entirely. Zuckerberg: “Do you want a future where you’re interacting with kind of one system for everything? Or do you want one where a lot of different people are building a lot of different AIs?” One AI controlled by one company. Or millions of AIs built by millions of people. Centralized intelligence. Or distributed intelligence. Zuckerberg chose distributed. Zuckerberg: “What open source does is it makes it so everyone can take and modify the model and build stuff on top of it. Which is different from the kind of closed and centralized approach.” The closed model makes every business a tenant. You rent intelligence on someone else’s terms. At someone else’s price. Inside someone else’s guardrails. The open model makes every business an owner. You modify the model. You deploy it your way. You build equity in your own system with every iteration. That gap widens quietly. Then it becomes permanent. The tenant pays more for less control every year. The owner pulls further ahead every cycle. One is a subscription. The other is infrastructure. Then Zuckerberg described the part most people have not thought about yet. Zuckerberg: “A lot of creators will have their own AIs. It’s like a richer world when there’s a diversity of different things.” Your favorite creator will have an AI trained on everything they have ever made. Available to millions of people simultaneously. Responding in real time while the creator sleeps. That is the difference between a brand that scales with your waking hours and one that scales with compute. One has a ceiling. The other does not. Zuckerberg is not betting on one model that governs everything. He is betting on billions of specialized AIs, each built by the person closest to the problem it solves. The companies still debating whether to adopt AI are not having the wrong conversation. They are standing in a room where the meeting ended an hour ago. The checklist updated. They did not.

Indian MSMEs run on WhatsApp, Excel, and trust. AI hasn't touched them. Yet. India has 63 million MSMEs. 31% of GDP. 250 million jobs. Ask any owner in Surat, Ludhiana, Tirupur, or Nagpur if they use AI in their business. Most will say yes. They mean WhatsApp. Or someone on their team opened ChatGPT once. That is not automation. That is not a workflow. That changes nothing about how the business actually runs. Real AI deployment, the kind where a process runs without a human triggering it, where data moves between systems automatically, where follow-ups go out without someone typing them, that is essentially at zero in Indian MSMEs. Not 7%. Not 2%. Essentially zero. Why this is the biggest untapped market in India right now. India's large enterprises are moving fast. 47% of them have AI running in production (EY-CII, 2025). Their MSME suppliers, distributors, and vendors? Still on Excel. Still on manual data entry. Still on phone calls to confirm orders. The gap between enterprise and MSME on AI is not a technology problem. It is a deployment problem. The tools exist. n8n, Make, Claude API, GPT-4, Zapier. All available. Most either free or under Rs 5,000 a month. What doesn't exist is a person who walks into the MSME, understands the workflow, and builds it. That person is the AI Workflow Architect. What this person actually does. Real example. A garment exporter in Tirupur processes 200 orders a week. Each order needs: Buyer email parsed PO data entered into Tally - Production schedule updated - Shipping documents generated - Buyer follow-up sent Currently: 2 data entry operators. 8 hours each. 5 days a week. - An AI Workflow Architect builds this in 4 weeks: - Email parser using Claude API or GPT-4 - Tally integration via API - Auto-generated shipping docs - WhatsApp follow-up bot Cost to client: Rs 2-3 lakh one-time. Rs 15,000 per month to maintain. Savings to client: Rs 40,000 per month in salaries. ROI in 6 months. This is not complicated. It is not being done because nobody is walking in to do it. The IT crisis and the MSME gap are the same story. Fresher IT hiring: 600,000 in FY22. Down to 120,000 by FY25. An 80% drop in three years. (Source: Xpheno) TCS cutting 12,000 jobs. NITI Aayog warns of 15-20 lakh IT jobs at risk. Everyone is looking at that number and panicking about what's ending. Nobody is looking at the 63 million businesses that need someone to deploy AI into their operations. The same disruption that kills the BPO seat creates the AI deployment market. These are not separate events. They are the same event, viewed from different angles. The skill set is learnable. In months, not years. - No CS degree needed. No advanced Python. - Prompt engineering learning time: 2 weeks - One automation platform like n8n or Make: 3-4 weeks - API basics, connecting tools to each other: 3-4 weeks - Reading a business process and mapping it: ongoing Three months of focused learning. Then you go find one MSME that has a painful manual process and you fix it. This is the time, this is the opportunity. India's future for next 3 decades will depend on this.

Given the PyPI supply chain attack, I recommend keeping a canary in the coalmine: I have a bitcoin private key containing $100 of BTC in my .bashrc. It's clearly labelled. If my system is ever compromised by some bad package, the BTC will get stolen, and I'll see the move on-chain. And that'll tell me that I need to rotate every single other secret. There are even services that will send you an alert (text, email, whatsapp...) if a given bitcoin address moves funds. It's good to have a burglar alarm, especially when time is of the essence.

Holy Moses he’s built different. If it’s easy, you didn’t go hard enough. Go til it hurts. Push. I only know a few ceos who run their business this way. And guess what. They’re winning.

Tell me a story that sounds fabricated but is 100% true. (Work edition) don’t air me 🥹

Basecamp founder Jason Fried: Customers don’t care about your features and technology Jason recounts how when he was selling shoes growing up in Illinois, representatives of the major shoe brands would arm salespeople with facts about all of the fancy new technology (e.g. the difference between the Nike Air versus the Zoom Air, the Goodyear rubber outsole and the midsole, etc.). However, when Jason told customers about these technologies, they didn’t care. When you actually watched customers buy shoes, they really only cared about a few key things: What does it look like? Is it comfortable? Can I afford it? Jason saw the same thing selling tennis rackets. He’d tell customers about all the latest technology and the difference between graphite versus fiberglass or natural gut string versus synthetic. Nobody cared. They’d look at a tennis racket and ask him if it came in other colors. Jason has observed the same pattern across website and software: “Companies are obsessed with features and the technologies because that’s what they do all day. But they don’t actually watch people buy stuff. When you watch people buy stuff, they just want the simple stuff. They want the stuff that solves their problems and just works. They like the way it looks. They like the way it feels. It’s comfortable. It’s affordable. That’s what people want…. Yet companies keep talking about the specs, the technology, and the features.” Jason advises founders to really listen to what your customers are saying and watch how they buy your product: “You’ll find out that they just need a few things done really, really well. And that is really, I think, the secret to all this stuff. It’s figuring out the basics. Nailing the basics.”