Shunya AI

  • Read what’s written in this piece by Bloomberg with patience. But remember, I always say – stick to first principles and natural laws. That’s what Einstein and some of the brightest inventors have done (not the money inventors).

    The questions you must ask are follows:

    1. Is this a rally before the correction/ crash?
    2. Who has already made money (actually a lot of them have – I missed the opportunity)
    3. Is this even a logical comparison to be made?

    By some estimates Indian households have about $3 Trillion worth of Gold reserves. Nithin tweeted about it:

    I am no expert at guessing. Actually we are all experts in this domain. Here’s my best guess: a lot of retail investors will get trapped in this rally. Especially the one’s that have experienced 30%+ returns on their Gold Mutual Fund investments. But then they will feel the shock when it breaks. And then they will pay taxes as per tax slabs.

    And then most finance journos will get back to hailing Hydra. I mean, hailing AI.

    Till then, use first principles and natural laws and make money.

    Did Gold Surpass AI – Bloomberg says so – for how long?

    –––––––

    Oct 7
  • I am not the smartest guy in the world. But I do believe in natural laws and first principles. They kind of almost never fail (to impress).

    The world of AI is booming/ hyping. Whatever. There’s this whole world that believes human intelligence is going to suffer and AI will take over. Ok. That’s what people are calling the future.

    The saints have taught us – live in the present. Let’s try it for once.

    The kind of capex investment going into AI and the kind of returns being generated is not sustainable in anyway. Okay, I don’t own a datacenter and I am not Nvidia or AMD or a Musk or an Altman. But as I am a first principles guy.

    The only thing I see going up are electricity bills in the near future. Unless someone cracks the code to sustained solar energy and then the world conversations shifts to solar and climate change again.

    There are stories of comparing this AI moment to the railroad moment. Railroads were a key necessity. May be I am missing something out. May be making humans more insanely efficient is also a key necessity.

    But then a safe assumption to make is: most people are leveraging AI in the most ‘dumb’ manner possible. To create content on social media. To write E-mails.

    I teach a course on applied AI and I tell my learners – folks, AI will only amplify who you are. So if you are intrinsically dumb, it will amplify that. If you are a critial thinker and strong at reasoning, it will amplify that.

    Why do I teach a course on AI when I am talking of the AI bubble? Ummm, I remember someone saying ‘in a Gold rush sell shovels’. I keep it honest. May be some of my learners will actually change their life or write the world future. But then anything can happen ‘may be’.

    I do appreciate the noise around agents and agentic ai. But apart from the buzzzz, I don’t see the bizzzz !

    May be I am not yet plugged into the AI matrix, but I see billions of dollars being spent on training machines to be smarter only to be used by someone a lot more dumber?

    This smart – dumb partnership is what makes me sound the bubble. Unless the user evolves to be way smarter than he/ she is today, AI will only be as useful. Or may be one day it will outsmart us all and rule the world.

    MONEY

    Whatever it be. Some people will make boatloads of money either selling shovels or investing in AI infrastructure companies. There’s a lot of life changing money at stake here.

    Look, AI, even if its a bubble, it will take another couple of years for the signal to come through the noise. And that’s when some people will emerge and most will see yet another hype world changing future shaping cycle.

    Now as I said, I may not be the most intelligent human. I might not have seen the cycles really closely. And I certainly haven’t made insane sums of money investing in any of these hype cycles. Infact I am selling shovels for the very first time.

    But I do see with clarity a new form of intelligence being leveraged by a large number of low intelligence beings. And if for some reason this is how AI gets trained, overall it will turn out to be the world’s worst hoch poch noodles.

    Remember folks, natural laws and first principles. Stick to them.

    The AI Bubble?

    –––––––

    Oct 7
  • The Networked State Conference shared valuable signals for anyone that wants to understand what’s happening in the world from a 30,000ft level. Watch the stream:

    China, Gold, Dollar and The Networked State – Conference

    –––––––

    Oct 3
  • I absolutely love the idea proposed by VR. The idea of UEO. People owning equity early on and becoming millionaires by the virtue of the fact that they invested right. So effectively everyone’s an investor. They are investing in a few doers / entrepreneurs that ultimately shape the future.

    Watch this interview:

    Universal Equity Ownership – Vivek Ramaswamy’s wealth creation proposal for the AI first world

    –––––––

    Oct 2
  • The date: 28th September 2025. The time: 8:29am. The day: Sunday.

    Its been 9 months to Shunya AI (operationally active). This post will capture the lessons and the future that we are looking to build.

    The start

    Jan 2025 – I started as a solo founder with a dream to build a global online school. A school where the youth will come to learn 3 core skills: AI, Mindset and Business.

    I deeply respect and appreciate the first 50 learners that trusted Shunya AI and joined our Batch 1. And then started Batch 2,3,4,5 and 6.

    Sep 2025 – we are 1300 learners from 15 countries of the world.

    The secret behind the growth?

    Honesty. That’s it. My videos on facebook went viral. I received 50M+ views. These views converted into sales.

    In July 2025 facebook did something to its algorithm and the views started going down drastically. Our reach was stifled. This clearly reflected as a drop in number of new admissions to the AI LABS course in July, Aug and then Sep.

    As a founder I of-course found this a stressful and confusing. I was doing exactly what I was doing before, nothing changed. Then why this drop? Facebook support team also found it amusing. They said “there are no restrictions on your page, we also find this strange”.

    Anyway, I learnt not to panic. I calmed myself and reflected on the vision to establish one of the top 3 most trusted online schools in the world for the youth and young entrepreneurs.

    I chose not to run any ads to promote our courses, even now.

    Wait, before we speak of the future

    The fact that we have 1300 learners from 15 countries, without any ad spend is an achievement unlocked. In the online Education industry, where every company is selling courses from INR 99 onwards, we maintained our position. AI LABS course fee started from INR 10,000. In Sep 2025 it is INR 12,500. We have improved the course content and the delivery too. With every iteration we improved, massively.

    Cut to Sep 2025

    We have just started Batch 6. The admissions metric is at its lowest (compared to the previous months).

    Today the question that I am trying to answer:

    In this noisy world of social media, how do we not run ads and still reach students, professionals and young entrepreneurs with trust.

    I did meet Ed-tech entrepreneurs. These are companies that are spending anywhere from INR 20 lakhs to INR 40 lakhs every month in paid ads. They are effectively buying attention, building funnels and trying to convert those funnels with sales teams or by hosting large free offline or online events.

    I am personally not tuned to this method of acquiring new customers.

    The Future

    1. Marketing: I am going to invest a large part of my time in Educating people online – Educating is going to be our marketing. This will also attract a better quality audience. We will improve the quality of our videos to make them look like they come from a trusted online school
    2. Sales: With data I am pretty convinced that hiring a sales team to answer customer queries does not create an exponential impact for an Ed-tech firm. I have personally taken 250+ calls till date. The quality of calls are incredibly poor
    3. At the product level – AI Labs is a course that I have seen maturing from Batch 1 to Batch 6. Even my kids learn from AI Labs. Should we launch more courses? Should we re-look pricing?
    4. Operations: Should we start an offline center? I am personally not too inspired by this thought.
    5. Business Plan/ Revenue: Is there a B2B play here? Should we partner with colleges and somehow run AI Labs across colleges in India and otherwise across the world?
    6. People: Is this the time to hire someone really good, someone much better than ‘I’ to lead marketing?

    As a founder

    I am definitely enjoying this journey. This journey will create a legacy that will outlast me.

    If you are someone inspired to shape the future by Educating the youth on AI, write to me: reachps001(at)gmail(dot)com

    Thank you for your time and gratitude folks. I wish you a great legacy.

    The Shunya AI journey and the vision till 2028

    –––––––

    Sep 28
  • This keynote from Sequoia AI Ascent 2025 is worth the watch.

    At Shunya AI our understanding is that there is too much relevance being given to AI agents. When reasoned logically, it seems quite obvious that AI agents can massively improve human efficiency. But we still feel, AI agents are counter intutitive to the way humans have evolved to work.

    Humans are by default designed and trained to work. A few people rise up to the maturity of managing work (than doing the work). That is why there are fewer very highly trained managers.

    We feel that the world will need more AI READY generalists that will manage agents.

    Of-course we could be very wrong. If AGI evolves, the way it seems it will, ideally AI will learn and do almost everything by itself. May be then humans will have to think their own relevance.

    Till then, learn from this keynote.

    Thank you Sequoia AI Ascent 2025 for bringing forth some key points in the AI evolution.

    AI’s Trillion-Dollar Opportunity

    –––––––

    May 8
  • Last evening the Shunya AI team met the Krutrim team at an event in Gurugram. We learnt about the progress that Krutrim has made over time.

    Krutrim Progress:

    1. Investment of INR 10,000 crores in setting up India’s largest GPU cluster
    2. India has been restricted by the US on how many GPUs we can get
    3. Krutrim team showcased the progress made across Krutrim 1,2 and the “3” that they are launching soon

    What excited us?

    First things first. A company that’s building in Bharat. For Bharat and then for the world. That in itself is a solid beginning. Of-course we are late.

    1. The use case of voice (Bharat is voice first and text second in our experience)
    2. We also liked how Krutrim is being trained on Bharat first contexts (in both images, dialects) etc.

    The opportunity that we saw

    1. Its great to build GPU clusters, LLMs and the great tech
    2. BUT – how will you make Bharat’s students, professionals and MSMEs be AI ready? ( भारत के युवा को कैसे पता चलेगा कि करना क्या है ? ये सारी बातें तो उनके लिए अनिल कुंबले कि गूगली हैं | )

    This is exactly the problem that we are solving at Shunya AI. We are training Bharat’s youth to be AI ready.

    The Krutrim event could have been better

    This event could have been a lot more powerful. We are talking Bharat. But this event lacked the Bharat energy.

    More power to Krutrim

    In Bharat we generally find gaps first, instead of supporting each other. In 2013 my first startup was accelerated by Intel & Berkeley in Silicon Valley. I learnt there first hand how the ecosystem supports you and that’s hwo you progress.

    At Shunya AI, we will support Krutrim. Our team has decided that we will partner with Krutrim and train Bharat’s youth on Krutrim models.

    This might seem like a small drop in the AI ocean. But in the longer run, this will have a critical and compounding impact for Bharat.

    Ola Krutrim – Bharat’s AI Movement?

    –––––––

    Feb 25
  • It is clear that the world is now divided in 2 parts.

    1. People that have adopted AI as a way of life
    2. People that are still thinking about what to do and how is AI going to impact them

    The 16 most critical AI skills that will prepare you to be the top 1% globally:

    1. Business as a skill
    2. How to crack entrance exams as a skill
    3. Strategy as a skill
    4. Investing as a skill
    5. Communication as a skill
    6. Spycraft / Tactic as a skill
    7. Invention as a skill
    8. Negotiation as a skill
    9. Extreme focus as a skill
    10. Social influence as a skill
    11. Health and fitness as a skill
    12. Extremely fast decision making as a skill
    13. Act like a CXO as a skill
    14. Public life as a skill
    15. Tackling stress and anxiety as a skill
    16. Fake narrative fact check as a skill
    17. Leverage time as a skill

    Who will benefit from these skills?

    1. Students
    2. Professionals
    3. Startup founders
    4. Small business owners

    Summary

    As you develop these AI First skills, you become part of the elite top 1%

    16 AI Skills To Be Top 1% Globally

    –––––––

    Feb 24
  • The AI industry is witnessing a paradigm shift. While large-scale players like OpenAI, Google DeepMind, and Anthropic are building foundational models—Large Language Models (LLMs) and Small Language Models (SLMs)—the real opportunity for entrepreneurs lies in creating Value-Added Services (VAS) on top of these models.

    The path to wealth in AI isn’t in competing with these giants but in leveraging their infrastructure to solve specific problems at scale.

    The VAS Opportunity in AI

    Foundational AI models provide raw intelligence, but they lack contextual specialization, domain expertise, and user-friendly workflows. Entrepreneurs can unlock massive value by building AI-driven solutions that tailor LLMs and SLMs to specific industries, workflows, and consumer needs.

    This approach has parallels with the early days of the internet. Amazon didn’t invent the internet, but it built an e-commerce empire on top of it. Similarly, Uber leveraged GPS and smartphones, but its real value came from streamlining ride-sharing.

    In AI, the winning formula is applying foundational models to niche problems with high-value solutions.

    Examples of Value-Added Services in AI

    1. Jasper AI – AI-Powered Content Creation

    Jasper AI built an AI-powered writing assistant by fine-tuning GPT models for marketing and copywriting. Instead of competing with OpenAI, they packaged AI-generated content as a service, focusing on ad copy, blog writing, and sales emails. Result? A $1.5 billion valuation within two years.

    2. Replit – AI Coding Assistant

    Replit took AI-assisted coding to another level by integrating AI into their online coding environment. By offering instant code completions, debugging, and collaborative features, they created a VAS that enhances software development workflows, making coding faster and more accessible.

    3. Harvey AI – AI for Legal Services

    Harvey AI built an AI legal assistant using GPT models, focusing on contract analysis, due diligence, and legal research. Instead of developing a new AI model, they optimized existing LLMs to understand legal terminology and nuances, securing partnerships with top law firms.

    4. Runway – AI for Video Editing

    Runway applied AI to video content creation by leveraging LLMs for scriptwriting and computer vision for AI-powered video editing and special effects. Their platform is now used by filmmakers, marketers, and content creators worldwide, showing how AI can disrupt creative industries.

    How to Build a VAS Business in AI

    1. Find a Niche with a Pain Point

    • Identify industries where AI can drastically improve productivity or reduce costs.
    • Example: Customer support automation, legal document analysis, or AI-driven medical diagnostics.

    2. Leverage Existing LLMs & SLMs

    • Don’t reinvent the wheel. Use APIs from OpenAI, Anthropic, Mistral, or Meta to build AI-powered applications.
    • Example: Notion AI integrates AI-generated writing directly into their productivity tool, rather than building their own model.

    3. Build Workflow-Specific AI Solutions

    • AI alone isn’t enough; users need seamless workflows.
    • Example: Zapier AI integrates LLMs into automation workflows, helping businesses streamline tasks effortlessly.

    4. Focus on UX & Trust

    • AI models are powerful, but without good UX and reliability, they won’t get adopted.
    • Example: Grammarly transformed language models into a simple, trustworthy writing assistant.

    5. Monetize via SaaS, API, or Enterprise Sales

    • Subscription Models: Charge for AI-powered tools as SaaS (e.g., Jasper AI, Runway AI).
    • API Access: Offer AI-driven capabilities via APIs for developers (e.g., ElevenLabs for AI-generated voices).
    • Enterprise Customization: Provide AI solutions tailored for businesses (e.g., Harvey AI for law firms).

    The Future of AI Entrepreneurship

    The AI revolution is still in its early stages. Foundational models will continue to improve, but the real impact will come from entrepreneurs who build specialized AI-driven services. Those who can combine domain expertise, user experience, and AI capabilities will create the next wave of billion-dollar businesses.

    If you’re an entrepreneur, the message is clear: You don’t need to build the AI model—you need to build on top of it.

    Are you working on an AI-powered VAS? Let’s discuss how to scale it to the next level! 🚀

    The AI Gold Rush: Building Value-Added Services on LLMs and SLMs

    –––––––

    Feb 18
  • Artificial Intelligence (AI) has revolutionized the world, and companies like OpenAI, Anthropic (Claude), and Perplexity are at the forefront of this transformation. However, none of these groundbreaking AI innovations originated in India.

    This raises an important question: why hasn’t India, with its vast talent pool and booming tech industry, produced an AI pioneer of this scale?

    The answer lies in the fundamental building blocks of AI itself—critical reasoning and the art of questioning—both of which are conspicuously absent in India’s education system.

    The Role of Critical Thinking in AI Development

    AI, at its core, is built on the ability to question, analyze, and infer. The foundation of AI systems like GPT or Claude is deeply tied to the art of asking meaningful questions and formulating responses based on logical reasoning.

    The greatest minds in AI didn’t just accept existing knowledge; they challenged it, experimented, and reshaped it.

    From Alan Turing’s foundational work on computation to modern deep learning breakthroughs, AI has always been driven by critical thinkers who push boundaries rather than conform to them.

    The key ingredients behind AI’s rapid evolution are:

    • Curiosity: The desire to ask better questions.
    • Critical Reasoning: The ability to challenge established norms.
    • Freedom to Experiment: The willingness to take intellectual risks.

    Unfortunately, these very attributes are discouraged in India’s education system.

    The Colonial Hangover of India’s Education System

    The Indian education system is still largely structured around rote learning, a relic of the colonial past where the British designed it to produce clerks and administrators rather than thinkers and innovators. Instead of fostering curiosity, it rewards memorization. Instead of encouraging debate, it punishes questioning.

    How the System Stifles Innovation

    Lack of Encouragement for Questioning: Indian students are often discouraged from challenging their teachers. The classroom remains a space where knowledge is dictated, not debated.

    Examination-Oriented Learning: Students are trained to score well in exams, not to think critically or solve real-world problems.

    Bureaucratic and Rigid Academic Culture: Research and innovation are stifled by red tape and lack of autonomy, making it difficult for disruptive technologies to emerge.

    Aversion to Risk and Failure: AI pioneers in the West embrace failure as a stepping stone. In India, failure is stigmatized, pushing students toward safe career paths like engineering and medicine rather than entrepreneurship or research.

    Outdated Curriculum: While the world moves toward AI-driven interdisciplinary learning, Indian universities still emphasize outdated syllabi with little practical application.

    Where India Excels but Falls Short

    India undoubtedly produces some of the world’s best engineers and computer scientists. IITs and other top institutions churn out brilliant minds who often excel when given opportunities abroad.

    Yet, the same individuals who fail to innovate within India go on to lead groundbreaking AI teams in companies like Google, Meta, and OpenAI.

    This is because the environment outside India nurtures risk-taking, intellectual autonomy, and critical discourse—things the Indian education system actively suppresses.

    The Need for an Educational Revolution

    If India truly aspires to be an AI leader, it must:

    • Reform its education system to prioritize critical thinking over rote learning.
    • Encourage questioning and debate in classrooms from an early age.
    • Foster a culture of research and risk-taking in universities and startups.
    • Reduce bureaucratic hurdles in scientific research and AI development.
    • Invest in AI-driven startups rather than just outsourcing talent to global tech giants.

    Conclusion

    India has the talent, the numbers, and the ambition. But without a fundamental shift in how it nurtures intellectual curiosity, it will continue to be a nation of AI consumers, not creators. The next OpenAI or Claude can come from India, but only when questioning is encouraged, critical thinking is prioritized, and innovation is not sacrificed at the altar of rigid educational traditions.

    Why Wasn’t OpenAI, Claude, or Perplexity Invented in India?

    –––––––

    Feb 14
Next Page

Copyright (C) 2025 – Shunya AI

 

Loading Comments...
 

    • Subscribe Subscribed
      • Shunya AI
      • Already have a WordPress.com account? Log in now.
      • Shunya AI
      • Subscribe Subscribed
      • Sign up
      • Log in
      • Report this content
      • View site in Reader
      • Manage subscriptions
      • Collapse this bar