I turned 14 this year. I also shipped a browser AI extension with persistent background agents, multi-tab swarms, six AI provider integrations, cross-app orchestration, and a workflow marketplace. People often ask me how those two facts coexist. This post is my attempt at an honest answer.
I am not writing this to be inspirational or to position myself as some kind of prodigy. I am writing it because I think the story behind Tensor matters for understanding what the product is and where it is going. The decisions we made — privacy-first, free, open, model-agnostic — are not just business strategy. They come from a specific worldview shaped by growing up during the most transformative period in AI history.
How It Started
I started programming when I was 10. Not because someone told me to, but because I wanted to build a mod for a game I was playing. The mod did not work. But the process of trying to make it work — reading documentation, debugging errors, asking questions on forums, staring at stack traces until they started making sense — hooked me. I realized that programming was not about memorizing syntax. It was about decomposing problems into pieces small enough to solve.
By the time I was 12, I was building Chrome extensions for myself. Small utilities — a tab manager, a bookmark organizer, a dark mode toggle. I built them because I needed them and because nobody had built exactly what I wanted. Chrome extensions were appealing because the feedback loop was instant: write code, reload extension, see results. No deploy pipelines, no infrastructure, no waiting. Just you and the browser.
Then ChatGPT launched, and everything changed. I started using it for homework help, then for coding assistance, then for brainstorming. But the workflow was clunky — switch to ChatGPT tab, paste context, get answer, switch back, apply answer. I thought: what if the AI lived inside the browser, right next to the content I was working with? What if it could see what I was seeing and act on it directly?
That question became Tensor.
MingLLM and the MIT Connection
I founded MingLLM as the company behind Tensor. The name comes from "Ming" (a family name) and "LLM" (large language models). It started as just me, writing code in my room after school. But as the project grew, I realized I needed guidance from people who understood both the technology and the business of AI.
Through a combination of cold emails, hackathon connections, and what I can only describe as persistent optimism, I connected with researchers and mentors affiliated with MIT's Schwarzman College of Computing. The College of Computing, which houses CSAIL (the Computer Science and Artificial Intelligence Laboratory) and connects with MIT Sloan School of Management, became a source of invaluable mentorship.
The researchers I connected with did not build Tensor for me. They challenged my assumptions, pointed me toward relevant research, and helped me think about problems at a deeper level. When I was building the swarm architecture, a CSAIL researcher helped me understand the distributed systems literature that informed our coordinator-worker pattern. When I was designing the privacy architecture, a conversation with someone in the Sloan orbit helped me think about privacy not just as a feature but as a competitive moat.
Being affiliated with MIT as a 14-year-old is unusual. I want to be honest about what it means: I am not a student there. I am not enrolled in any program. I am a young founder who has been fortunate enough to build relationships with people in that ecosystem who believe in what I am doing and are generous with their time and expertise. I mention the affiliation because it is real and it matters, but I do not want to overstate it.
What It Is Actually Like
Building a company at 14 is weird. There is no playbook for it. I cannot sign contracts without a parent. I cannot open a business bank account. I cannot attend industry conferences without someone accompanying me. When I join video calls with potential partners or advisors, there is always a moment of surprise when they see my face.
School is a constant negotiation. I attend school full-time, and my grades matter to me. Tensor development happens in the early mornings before school, evenings after homework, weekends, and school breaks. During crunch periods — before major releases — sleep suffers. I am not going to pretend it is sustainable. It is not. But it is temporary, and the work matters to me enough to make the trade-off.
The hardest part is not the coding or the time management. It is being taken seriously. When a 14-year-old says "I built a browser AI extension with persistent agents and swarm architecture," the default assumption is that it is a school project, a toy, or that someone else actually did the work. I have learned that the only way to overcome this is to let the product speak for itself. Download Tensor. Use it. Judge it on its capabilities, not on the age of the person who built it.
Why Privacy-First, Why Free
Some of the core decisions behind Tensor come directly from being 14. I am part of the first generation that grew up with social media from birth. I have watched my peers share everything online, often without understanding the consequences. I have seen how data collection, targeted advertising, and algorithmic manipulation work. And I have seen how hard it is to opt out once you are in.
When I designed Tensor's architecture, making it privacy-first was not a business decision. It was a moral one. I do not want to build a company that profits from surveillance. I do not want to be the person who built a tool that reads your emails and sells the insights. There are enough companies doing that. The world needs alternatives, and I wanted Tensor to be one.
Making Tensor free came from a similar place. I am 14. I do not have a credit card. Most of the tools I use, I use because they are free. I know that the best browser AI tool should be accessible to everyone — students, people in developing countries, hobbyists, and anyone who cannot justify a $20/month subscription. You bring your own API key, you pay your provider directly for what you use, and Tensor stays free forever.
What Drives Me
I believe the browser is the most important application on every computer. It is where people work, shop, learn, communicate, and entertain themselves. Yet the browser itself has barely evolved in 15 years. Tabs, address bar, bookmarks, extensions. The fundamental interaction model has not changed since Chrome launched in 2008.
AI changes that. With AI, the browser becomes an active participant in your workflow rather than a passive window to the web. It can understand what you are doing, anticipate what you need, and take action on your behalf. But only if the AI is built right — private, reliable, fast, and respectful of your autonomy.
That is what drives me. Not the novelty of being a young founder, not the desire to build a billion-dollar company, but the genuine belief that browser AI done right will make millions of people's daily lives meaningfully better. Every person who saves 30 minutes a day on repetitive browsing tasks gets that time back for things that matter. Over a year, that is 180 hours. That is almost eight full days. For millions of people.
What Is Next
MingLLM is growing. I am building out a small team of contributors who share the vision. We are expanding Tensor's capabilities, growing the workflow marketplace community, and exploring new frontiers in browser AI that I am not ready to announce yet.
I am also still in school. I still do homework. I still play video games (less than I used to). I still argue with my parents about screen time, which is ironic given that my screen time is literally building a company.
If you are a young person reading this and wondering whether you can build something real: you can. The barriers to building software have never been lower. The tools are free, the knowledge is online, and the platforms are open. You do not need permission, a degree, or investors. You need an idea, the willingness to learn, and the stubbornness to keep going when things break. Things will break constantly. Keep going.
And if you are not a young person — if you are a seasoned developer or entrepreneur reading this — I hope you will judge Tensor on its merits. Download it. Break it. Tell me what is wrong with it. I would rather have honest criticism from an experienced developer than polite encouragement that does not help me improve the product.
Tensor is not a school project. It is the beginning of what I want to spend my life building: AI tools that respect their users and make the browser genuinely intelligent. I am 14, and I am just getting started.