# We're All Playing a Zero-Sum Game
lastEditedTime: 2026-07-17T13:59:00.000Z
createdTime: 2026-07-16T14:09:00.000Z
categories: ["AI","Startups","Insdr link"]
language: en
description: A reflection on all of us building. But only OpenAi and Anthropic make the $$
cleanedContent: I’m Guhan Srinivasan, a 24-year-old running a technology solutions business from Bangalore. I recorded these thoughts in Cairo on June 16.

## The zero-sum cycle

The AI market increasingly feels like a zero-sum game. Nearly everyone is building a tool, but relatively few people have customers. Builders invest time, money, and tokens into products while worrying that the next OpenAI or Anthropic release could make those products redundant.

Sales can outperform building in the short term, but even successful sellers have to ask how long that advantage will last. Consider AI tutoring: someone who struggles with mathematics can already ask ChatGPT a question, switch to voice mode, and receive a useful explanation. It is tempting to turn that experience into a specialized SaaS product, but the product may only attract a handful of customers before the underlying platform improves enough to replace it.

## Connect expertise with real problems

AI-driven layoffs are painful, but there is an opportunity in connecting people who need practical solutions with people who know how to configure existing tools. This is different from spending months building and marketing another generic SaaS product.

If a business needs an interface for its data or a retrieval pipeline for a video library, the best answer may not be a new production-grade platform. An expert could configure an existing open-source system, connect an LLM wiki, and set up an agentic workflow. The business gets a maintainable solution without paying unnecessary development and API costs.

Only a small percentage of people understand these AI systems deeply, while a much larger group operates traditional businesses and needs help adopting them. Connecting those groups can create more value than repeatedly packaging the same capabilities into new subscription tools.

## A service built around relationships

Tools such as Wispr Flow, ChatGPT, and Cognee are becoming foundational. The opportunity is to teach people how to apply a small set of strong tools to their actual business and enterprise problems.

That is the idea behind **Insdr Link**: connect people who have business problems with people who understand how these systems work. Instead of building speculative products for little or no return, experts can spend a few focused hours with a customer, charge fairly for their time, and leave the customer with a sustainable solution.

The goal is not to sell another tool that might solve one narrow problem. It is to build relationships between people who need help and people capable of solving many problems over time. That creates useful work, keeps knowledge moving between people, and gives both sides a better alternative to the zero-sum cycle.
slug: were-all-playing-a-zero-sum-game
image: []
status: published
-------------------------------------------

It's 16th of June, 3:52 p.m. Cairo, Egypt. I’m Guhan Srinivasan im 24 I run a tech solutions company  out of Bangalore.


We are all playing the perfect no-sum game. I say this because the more time I spend in the market, the more I realize that everyone has a tool. A very small percentage of us have customers but everyone inherently is shitting themselves about the fact that what if the next update from open AI makes me redundant. This, in essence, means we pour a lot of money, a lot of effort, a lot of token usage into building a tool, thinking there's a problem that we've identified and we can solve. Most of us are good at building. The people who know how to sell are able to make more money than the people who end up just building. In essence even the people who do end up selling are uncertain about the fact that how long can we keep up these sales? I see this because now when I figure out a way that I can use ChatGPT to solve one of my problems. For example I'm studying and bad at math. The YouTube videos aren't helping me all that much so I put the question into ChatGPT. I read it and I still don't understand it that much so then I buy the premium and I turn on the voice chat and it does an excellent job at explaining it to me. The entrepreneur in me thinks, let's make a SaaS out of this, let's make a tutoring company. The way I came up with it is I used ChatGPT and I figured out, "Oh this is what works." If I build a voice agent that's a tutor where I'm niching down to solve my problem and building a tool, sure I'd get one or two customers but in the long run how sustainable is that? Because the more people realize how helpful this is, the more people will be pushed towards finding a better alternative to the solution that you've built.


Number two, the mass layoffs that are happening because of AI coding are unfortunate but we had that coming for a very long time. In my opinion, given how startups and the gig economy and all of that are advancing in my own city of Bangalore, I do think I see a way forward. That would be to try and connect the people who want these tools set up and the people who know how to set up these tools. By that I don't mean build out a tool, market it for 50 days, hope you make a few bits of money here and there. That is still a net loss because you're spending on API costs and coding out the tool. You're spending a lot of time marketing it and there's a 50- to 100-day turnaround time when everyone's getting laid off. They're all coming into this thinking that this is how we make money quickly. But at the same time this doesn't really solve the other person's problem. If someone needs an interface to arrange their data, they need a RAG pipeline or they need something like this. What's pushed into their faces is heavily funded sub-par open source software that was then combined and made to be the backend of a proprietary SaaS that works half as well as it would work if you just teach the other person how to use the open source tool for a fraction of the API costs for them. I do think what I see is this example in a billion different places. I see it if someone wants to make a RAG pipeline and they want to make it a production-grade RAG pipeline. It's going to take you years to get that set up. If you have to solve somebody's problem who has to use a RAG pipeline, for example, they're building a RAG pipeline for YouTube videos:

- Hook up a version of LLM wiki for them.
- Set up an agentic harness.

They have what they needed. You didn't have to spend two to three months developing a production-grade RAG pipeline to solve this Sumanth's problem and that's 10 people at max. What we could have done better there is if we were able to connect the people who have the know-how of how to build these things and we were able to connect the people who need these things. I think it's pretty widely known that the people who use AI were less than 10% of the world at the moment. The people who are operating in mass businesses are the bulk of us. I'd say 80-90% of the world operates business as usual. If you can connect the 10% who know too much about AI with the 70% who need AI but are a bit daunted to get into it, you can provide each other the simplest solution catered to them. Instead of saying, "Listen, I've built a RAG pipeline tool. Listen, I've built a tutoring tool. Listen, I've built a tool that brings you sales," or, if expertise and time are treated with the right payment structure, I think people will end up making a lot more money rather than us burning API costs and making anthropic and open AI and whoever is providing our stupid API money.


So as I say I use Wispr Flow here to get this thing set up. At the end of the day there are tools that people will end up using quite well, like Wispr Flow, ChatGPT, and Cogni. I think these are foundational tools but if we teach people how to use a few of these foundational tools to solve business problems, enterprise problems, I think people will extract a lot more value from it. I think economically the cycle of money would be better because it's not half big enterprise switching all the way to API costs and "let's lay off my engineers now." That's not sustainable. People need jobs at the end of the day. People need money. Money has to flow within people. With that said I'm not here to pitch a tool but I'm here to pitch a relationship, a service to the people who need money and a service to the people who need their problems solved. I'm calling it Insdr Link. This is particularly because the word "insider" in the domain was taken so I'm calling it Insdr Link, so insdrlink.com.


In essence what it's going to do is connect the people who have business problems with the people who know how these things work and who clearly have way too much time on their hands. This is because all we do is sit around and build tools for two, three, or four hours for little to no money.


How about you sit on a call with the people who need their problems solved for two, three, or four hours and charge, I don't know, $50 to $60 an hour? Make sure the solution you provide to the other person sitting across from you is something that they can use sustainably that's best for their problem forever. You're not trying to fleece them because you're getting paid anyway and two hours from now you have another meeting with another person who's looking to do the same thing. We're here to build relationships. We're not here to build tools that might solve a person's problem. We're here to introduce people to other people who can sustainably solve a lot of the problems that they're having and of course we’d monetise it. call it were all playing a zero sum game