I learned from building Shaadi.com that Indians pay when the outcome becomes important enough.
I described customer workarounds this way: We share subscriptions. Stretch free trials. Find jugaad workarounds.
I framed the central customer question as: Bhai, isse mera kya hoga?
I see India as a demanding market: India is not a cheaper version of America. It is a far more demanding market. Customers can smell ‘chep raha hai’ from a km away.
I think India-specific AI plans are teaching companies about value: As AI companies roll out India-specific plans, I think they are being schooled in value.
My view is: If an AI product can become useful, affordable and habit-forming here, It will become a better product everywhere.
I see India’s earlier impact on telecom, payments, and entertainment as a precedent for AI: India did this to telecom, payments and entertainment. Now AI gets its turn.
This pricing perspective builds on my broader view of how India reshapes technology and products for local needs.
I see distribution and local relevance as advantages in India: “Distribution beats brute force. Local beats generic. Always has.”
I see a gap between building another ChatGPT and building for Indian users: “Few are building the people’s AI for India.”
I tried Puch AI and described it this way: “Very colloquial. Very Indian. All languages. Use cases most global models don’t even think about.”
I tried a few local GPTs including Puch AI. The post describes the experience as: “Tried a few local GPTs including Puch AI. Very colloquial. Very Indian.”
My view is that “India doesn’t just consume platforms.” “It reshapes them 🇮🇳🤖”
I recalled that “Early 2000s, our biggest bottleneck at Shaadi.com wasn't engineering.” The bottleneck was photos, not engineering.
In the early Shaadi.com photo-upload process, “Half gave up before they finished.”
I said, “The photo decided everything before the bio got read.”
I noted that “Everyone has a 50MP camera in their pocket.” Even with thousands of photos, I described finding a suitable professional image as a problem: “Hazaar photos in the gallery, none of them quite right.”
I attributed the professional-photo statistic to LinkedIn: “If your LinkedIn starts pulling 21x more eyes, the compounding begins right there.”
I described the headshot process this way: “ChatGPT Images 2.0 does it in 5 seconds. Upload a regular photo, tell it what you want, done.”
I rejected simplistic headlines about India and AI, saying, “Neither is particularly useful.”
My view is: “India doesn’t need to win the model race to be in the AI race.” I argued that India’s opportunity includes enterprise AI products and services as models become abundant and the real battle shifts elsewhere.
I described Superleap AI CRM as attacking a ~$120 billion CRM market. “In under 2 years, it is already clocking multi-million $ARR, with sales cycles of just 3 months.”
I described Mitigata as targeting a ~$200 billion cybersecurity market. “In < 3 years, it has 650+ clients & revenues in 100s of cr this year.”
I cited Tata Consultancy Services as an example of established Indian IT benefiting from AI: “are now seeing nearly 10% of revenue come from AI-led work.”
I argued that enterprise AI products and services could help take India’s IT industry to ~$1 trillion over the next decade, while noting that the world is not close to any Singularity.
Across these posts, I connect local distribution and context-aware products with India’s opportunity in AI, from consumer-facing tools to enterprise products and services. I also see AI image generation as a practical application for the professional-photo problem, distinct from image tools used mainly for entertainment.
I invited people to build foundational AI models for India and said the world needs more AI platforms beyond big tech; it is India’s time. I framed this as an invitation to build for India and the world.
I said, “While 🇮🇳 hasn’t done it yet,” and expressed hope that India’s time will come. My post also recalled @sama’s words to India in 2023; those words belong to @sama, not me.
I asked who was building an Indian product using LLMs for better discovery. I did not specify what the shortened link referred to or identify a particular product category.
I asked who was building the Alibaba.com of India using LLMs for better discovery, and invited readers to tag someone they knew. I did not name a builder or describe an implementation.
I expressed confidence that India’s time would come, writing, “Apna time bhi aayega.”
I invited readers to tag someone who might be building an Alibaba.com of India with LLM-powered discovery.
Related: finding and validating opportunities, digital sovereignty and dependence, conviction against convention