Krishna Sahu

If you remove all the stocks that have gone up and then compare Indian market cap to Taiwanese market cap, we're still bigger. #Meths

Debates about gig work in India are often devoid of nuance, context, and reality. In fact, most of the people arguing about gig work probably don’t understand much about the real India. Take it from somebody who comes from a part of India that would not even be considered rural. Where I come from, earning ₹150 an hour would be considered an enormous luxury. And this is pretty much the story of the vast majority of Indians. People would kill to make ₹150 an hour. Whenever people argue about the sad nature of gig work, the brutal working conditions, the lack of social safety nets, the tracking, and everything else, the problem is that these debates lose all nuance. Yes, there is truth to many of these criticisms. Yes, these are all important issues. But you also have to realize that India is fundamentally a poor country where there are absolutely no employment opportunities for the vast majority of Indians who come from the rural heartlands and hinterlands. For many of them, getting a gig job is akin to getting a software developer job. Remember, these are people who go through life bearing the enormous burdens of their social circumstances, life circumstances, and social realities. They have to put food on the table. They have to pay the medical bills of their loved ones. They have to send their kids to whatever school they can afford. They have to keep a roof over their heads. For most Indians, this is the very definition of luxury. This is not to excuse all the shenanigans pulled by gig companies. This is not an apology for the lack of social safety nets. None of that is what I’m saying. These are all important issues, and they are nightmarish issues. There are no easy answers. There is no black and white here. But to say that all gig jobs are bad, to scoff and sneer at gig work, and to take positions that are not rooted in lived realities without speaking to the people actually doing this work, always gets on my nerves. Whenever I speak to an Uber driver, or whenever I use services like Urban Company, the one thing I ask them is: how are the working conditions? And very rarely do I hear complaints from these people. Of course, I work in finance, and I know that the plural of anecdote is not data. There are people in the gig sector who work under brutal, punishing conditions for a pittance. But you also have to realize that a lot of these people don’t have any alternatives. Of course, we have to fight to ensure that their working conditions improve. We have to speak up for better protections, better pay, and better safety nets. But that is a continuous journey. So, to people in urban cities sitting under the comfortable breeze of your Daikin AC, typing on the latest iPhone, wearing your Birkenstock slippers, drinking your ₹300 matcha latte: go spend a day in the real India. Go speak to actual gig workers. Spend a day in their lives. Trust me, I’ve done it. I’ve done gig jobs. I come from a poor part of India. A lot of people from my town are employed as gig workers. So I speak from having been on both sides of this debate. The vast majority of these people would choose a gig job over being unemployed and unable to take care of their loved ones. That doesn’t make gig work perfect. It doesn’t make the exploitation acceptable. It doesn’t mean companies should get a free pass. But it does mean that if you want to have an honest debate about gig work in India, start with the lives of the people doing the work. Not with your ideological priors.

Let me translate that for you. FM: Look, we always listen to everybody. We always respect other people's opinions regardless of whether they are right or wrong. We are consultative and we always are open to feedback. Then we ignore it all and do what we want, data and evidence be damned.

Fuck Citadel, Jane Street, Millennium, Renaissance! Fucking $12 Claude prompts masquerading as "alpha" generating "hedge funds" Finally the scam has been exposed. If you have Claude, you can make Jim Simmons' ghost your unwitting bitch!

Always own beta. No matter what you do, always own beta. How much beta you own is an entirely different question, but you should always own some beta. One of the first things I and many of my colleagues learned, after getting our teeth bashed in by the markets and paying a high price for it, is this: always own beta. All the other fancy things, all the high-risk, high-reward things, all the other strategies, everything else comes later. Beta is at the heart of everything. You would be surprised how many otherwise smart and thoughtful people keep learning the same lesson over and over again. They ignore beta, and then they get their metaphorical teeth bashed in. At this point in my life, having spent enough time in the markets, having seen different parts of the buy side and the sell side, this is one lesson I’ve truly internalized. So much so that it’s tattooed on my brain: Always own beta.

If you use markov chain models to lose money, it feels more effortful and satisfying. The goal in the stock market is not to make money. That's not possible. The goal should be to lose money in a way that makes you feel like you did something.

The glee with which people were dunking on Benn’s decision to shut down the fund was sickening and disgusting. I was lucky enough to get a shot at being in the market very early in my career, and I’ve been here for a couple of decades now. The one thing the market repeatedly teaches you is humility. It doesn’t matter how smart you are. It doesn’t matter how talented you are. It doesn’t really matter how much money you have. Everybody goes through a phase where they end up on the bad side of the market. That’s just the nature of being in the market, whether as an individual trader or as someone who manages outside money. And whenever such things happen, the wrong lesson to draw is, “Oh, that person was bad at managing money” or “He did something wrong.” The real lesson is that making money in markets, whether in an individual capacity or as a fund manager, is a blood sport. It’s akin to being a gladiator of yore in a Roman coliseum, trying to earn one’s freedom. And here, freedom is a metaphor for making enough money to put food on the table. Put another way, managing money is exactly like trying to build a business. It’s power law all the way down. It’s incredibly hard. It takes guts to show up and try to build something despite knowing the odds. And that’s exactly what people like Benn do in the market. They take on an insane amount of risk and try to build something against the odds. For that, you have to admire their courage and fortitude. A lot of things are rarely in our control in the market. I know it’s clichéd to quote the Stoic philosophers in an age where their entire message has been taken, chewed up, digested, pre-processed, and bastardized into cute little sound bites. But knowing what’s in your control and knowing what’s not in your control is essential in the market. There are always periods like this, like the one Benn encountered, where things just go south and there’s very little you can do. I have a personal story. My own dad had built a successful business, and the 2008 crisis, coupled with some regional political instability, came out of the blue and decimated him. He had to shut down his business and went through what was easily one of the most horrendous periods of his life. He struggled to put food on the table. But despite that, somehow, he picked himself up, dusted himself off, and ensured that his family didn’t suffer too much. So yeah, being in this business is an incredibly humbling experience. You may agree with a person’s views or disagree with them, but this gleeful joy, this schadenfreude people have at other people’s failure, is just nauseating. We are on a platform like Twitter where this is the default incentive and default impulse, thanks to the shape of the platform. It’s sickening. We all go through life. We rise up and we fall down. It’s better to always cheer for people’s success than celebrate their failure. The true mark of being really human is whether you can lend a hand or cheer people on, no matter what. I also want to express a profound amount of gratitude to Benn Eifert for everything he has shared. As someone who has picked up an insane amount of learning from his tweets, a simple thank you wouldn’t suffice, but that’s all I can do at this moment. The post on Noah Smith’s blog is one of my favorites, and I keep sharing it with the arrogant idiots in this industry. And I was a retail investor. So, Benn, thank you for all that you have done. This is just a bad phase. I’m 100% sure you’ll bounce back stronger than ever. Also, I’m really sorry about all the idiotic piling on you had to see on your feed.

A few thoughts on the AI slop debates. Go look at any platform. The slop videos tend to have a hundred times the views of other videos. I don’t know if you’ve been paying attention to this garbage on Instagram, where people are watching random cat videos, random emojis, and god knows what other brain-rot shit, but this stuff is racking up millions of views. A recent example that must have popped up on your feed was that Times of India article that was almost entirely AI-written. Then there was the recent Granta Prize article that was AI-generated. I think this is also just a side effect of a larger confusion. There are two different things happening here. One is outright AI-generated slop. Mindless, one-shot goop. The output of lazy, thoughtless prompting, explicitly designed for the industrial-scale generation of content. Industrial-scale generation of output. There is no thought here. No craft. No thinking. No effort. But then, in my personal experience, there are an insane number of writers who are using AI as a partner in their creation. And this new chimera-like partnership is leading to some very weird and very wonderful forms of writing, art, music, and other artistic artifacts. I’m personally a huge fan of this. Right now, “AI slop” has become a catch-all term for anything we don’t like. I think that’s a problem. Whether we realize it or not, AI is here to stay. Whether we like it or not, it’s here to stay. And I’m finding it hard to make peace with the fact that we are in this weird uncanny valley where any artistic expression, or any expression even remotely touched by AI, immediately becomes suspect. It becomes a defect. I think that’s unfair to a lot of creators and writers who use AI to come up with new forms of expression or to enhance their output and capabilities. I fall into this bucket too, although I’m neither artistic nor creative. I’m just some random fellow on the internet. And if I think about my own reading diet over the last six months, some of the best articles I’ve read, especially in places like Substack, have been explicitly labeled as AI-assisted writing. The authors have been honest enough to say so. And the pieces were really good. Really insightful. For example, the experiments run by Venkatesh Rao at Ribbonfarm and It’s a Trap come to mind. We are heading into this weird, jarring, disorienting world where AI wraps itself like an exoskeleton around people, enabling them to do more. In such a world, we need new labels and new categories for creative output and media output. This isn’t an apology for outright AI slop. In fact, I wouldn’t even call it slop, because slop is at least food that animals eat. This stuff is AI slime. It is the equivalent of a torrent of effluent and sewage flowing downstream with no end and no purpose. It takes a monumental amount of effort to repurpose it into anything useful. But we still need to distinguish between two very different things. On one hand, there is AI-assisted creative and media output where a lot of effort, thought, planning, and actual thinking has gone into wrangling and leveraging AI tools to create new forms of expression and new frames. On the other hand, there is the mindless AI slime that is exquisitely designed for nothing other than gaining clicks, views, or whatever other idiotic metric people are chasing. This suspicion around AI-generated slime is also a case for being more anal about what one consumes. Funnily enough, among a lot of thoughtful people, people who read and watch a lot of content voraciously, suspicion has become the default impulse. A lot of people now start with the premise that everything is AI-generated and garbage. Then it is on the writer or creator to prove that it is not. And by AI-generated here, I obviously mean slime. Not well-thought-out work that’s AI-assisted. Not work where AI is part of the process. I mean the mindless, synthetic, industrial sludge that is flooding all platforms. This is an uneasy world for creatives. Writers, artists, musicians, filmmakers, designers, journalists, creators, all of them are now operating in an environment where the audience is not just asking, “Is this good?” but also, “Was this made by a human?” or worse, “How much of this is slop?” Calling all creative artifacts “AI slop” is lazy. It collapses a useful distinction. More importantly, it blinds us to the strange and wonderful forms of expression that are already emerging from this human-AI partnership. This new world, whether we like it or not, is here to stay. There’s no going back. From the slop, slime, and sewage will and must emerge the new and the wonderful. The real distinction is not AI versus human; that framing is already broken and counterproductive. The distinction is between thought and no thought. Between craft and no craft. Between intentionality and industrial-scale sewage. Between people using AI as an aid to think and create and morons using it as a machine to spray slime and sewage across the internet. We need better labels for this. Otherwise we’ll end up despising the tool, missing the art, and still drowning in the slime.

Apologies. If you make money, you're an options seminar seller. As Mahatma Gandhi once said, those who can't do teach.

I love Google. I really do. I think Gemini Flash, and even Gemini 3 Flash, is underrated in many ways. Used properly, especially through the API, Gemini models can be much better than a lot of other frontier models for many practical use cases. But goddamn, Google is terrible at messaging its AI products. The fucking jumble overlapping products is a nightmare. Compared to ChatGPT, the whole Google AI ecosystem feels like a jungle. There are so many front-facing products and experiments now: Pomelli, Stitch, Opal, Flow, Whisk, Mixboard, NotebookLM, Illuminate, Jules, Gemini CLI, AI Studio, Vertex AI, Gemini in Workspace, and probably a dozen more I’m forgetting. Some of these products are genuinely interesting. Pomelli is for small businesses and marketing campaigns. Stitch helps turn prompts into UI designs. Opal is a no-code app or workflow builder. Flow is for AI filmmaking. Whisk is for visual ideation. Mixboard is for concepting. NotebookLM is for research and working with sources. Jules is for coding tasks. Gemini CLI is for bringing Gemini into the terminal. But notice the overlap? The main model itself is capable of all of these things. Why create individual websites, apps, and flows in the first place? How will people discover this? These things are not even on the main app, even as links. Individually, a lot of these things make sense. But together, the product story is a mess. Why do Gemini CLI and Jules both need to exist in the same mental universe? Where does the Gemini app end and NotebookLM begin for research use cases? Why are some things Google Labs experiments, some things Gemini products, some things DeepMind products, some things Workspace products, and some things Cloud products? I’m sure there are good internal reasons for all of this. But from the outside, the messaging fucking sucks. The frustrating part is that the underlying technology is good. The Gemini API is outright underrated. If you use it the right way, it is fast, capable, cheap in many cases, and genuinely useful. But the consumer-facing product experience and the broader product positioning doesn't communicate that. Instead, Google keeps ejaculating one more named thing into an already crowded drawer. That’s the real problem. The models are good. Some of the products are good. The experiments are genuinely interesting but the branding, positioning, and messaging around the whole AI ecosystem confuses the hell out of me. I hope they get their shit together, because they have a lot of underrated products.

You'd be surprised at the number of humans who suck at this.

"Some 40% of Indians live in urban areas but they produce nearly 60% of GDP. In return they get broken footpaths, potholed roads, poor public transport, unbreathable air, sewage-clogged waterways and piles of festering rubbish. Imagine what they might be capable of if their towns were liveable." — Economist economist.com/asia/2026/05/1…

I don't know if India is 10 months away from becoming Somalia, and neither does anybody. Whether the critiques about the Indian economy are right or wrong, it's very important to remember what Dave Chappelle once said: Dave Chappelle: "Apparently they dragged me on Twitter. I don't give a fuck, because Twitter is not a real place." youtube.com/shorts/BlX6eo4…

My hot take on AI for normies open.substack.com/pub/sahukrishn…

@SachaSucha My Read It Later apps are a graveyard of links and wishlists. 🥲

@unrealyields Well said. I agree with literally everything here. The last points about the goop is spot on. You really should trademark that thought lol. Although personally I have successfully fought off the urge to "collect-em-all", but I still sometimes feel the urge.

I think entertainment works as an outlet to decompress only when there is slack elsewhere in the system. Because of economic realities, the structure of our, let’s say, post-2010 work life offers very little slack for people to just sit idly, stare at the clouds as they gently pass by. In my opinion, this is really, really important. Increasingly, it feels like a lot of our life is like driving a car at 100 kilometers an hour, in top gear, always. Which means you’re just accumulating a lot of gunk in your head that you otherwise don’t process. Call it the debris of life. That’s the first thing. The second thing, I think, is the degradation of all sorts of cultural artifacts and output. This has been spoken to death. They have increasingly been tailored, or rather molded, by those invisible potters we call algorithms. It has sort of become a loop, and I don’t want to make one of those lazy statements and say, “Oh, the algorithms are ruining everything.” That’s not what I mean. It’s just that this feedback loop of our preferences shaping the algorithm, and the algorithm’s preferences shaping us, has created this infinite regress. It feels like we are circling down the drain, and the quality of everything you see — TV, movies, writing, videos, podcasts, everything — has become sloppy in large part because we are pandering to the invisible mandate and dictum of the algorithms. This feels like a very tentative point, and I am not too sure about it, but maybe there is something to it: the sheer abundance of choice is making us dizzy and disoriented. Now we are more worried about the fear of not watching something versus the pleasure of watching something. Maybe I should trademark this. Then there is also the fact that there is increasing homogenization. This goes back to my point about cultural artifacts. There is increasing homogenization of everything. Everything just feels like a bland pastiche. It’s just gooped. Everything blends into everything else, and everything looks the same, feels the same, which means there is very little distinction. Maybe this is feeding into a little disillusionment. Again, this is one of the points that I’m not really sure about.

People in the United States are worrying that they will be part of the permanent underclass after AI. But if you are a middle class Indian you been part of the permanent under class all through your life. I have a post on this coming.

The second point gets people into more trouble than most people realize. I've seen so many allocators make the bet that the markets are wrong and get hit in the nuts so hard they escape through their mouths. Also, please follow @G3ni3sWish I always learn something from his tweets

How many Fintech apps does this kill?

Hello world.