What the Machine Sees
CN_000 — The Charaka Notes Manifesto
The average failed startup survives 4.2 years. Most founders never see the pattern that kills them — because no human can hold 13,499 companies in working memory. A machine can.
companies across 48 sectors, tracked in real time
The Problem We Couldn’t Unsee
We started building Manthan to solve a selfish problem: we wanted to invest better. But the deeper we went, the more we realised that the Indian startup ecosystem is flying blind.
There is no shortage of opinion. There is a catastrophic shortage of pattern.
When you track 13,499 companies across 48 sectors, digest 188 postmortems, and map the behaviour of 1,747 investors, something shifts. You stop seeing individual stories. You start seeing mechanics.
- Why do 34% of startups die from competition — not product-market fit?
- Why does the median failed startup survive exactly 4.2 years before collapsing?
- Why do certain investor combinations predict success 3.7x better than others?
- Why do regulation-killed companies take 5.1 years to die — longer than any other cause?
These aren’t opinions. They’re patterns. And they repeat with mechanical precision.
Why We’re Publishing This
In ancient India, Charaka was the physician who didn’t just treat patients — he catalogued disease itself. He turned centuries of scattered observation into the Charaka Samhita, a systematic framework that let any practitioner diagnose what they were seeing.
That’s what we’re building for startups.
Manthan’s AI doesn’t predict the future. It diagnoses the present — with enough depth and cross-referencing that the future becomes significantly less surprising.
We named this publication Charaka Notes because the ambition is the same: turn pattern into practice.
What You’ll Get
One insight every weekday. 500-800 words. Free. No fluff.
Each note draws from the full knowledge graph — 13,499 companies, 188 postmortems, 1,747 investors, 48 sectors — and surfaces one pattern worth knowing.
| Day | Pillar | What It Covers |
|---|---|---|
| Monday | Pattern Intelligence | Cross-sector patterns the machine detects — survival curves, timing mechanics, competitive dynamics |
| Tuesday | Sector Deep Dive | One sector, dissected: who’s winning, who’s dying, and the structural reasons why |
| Wednesday | Death Diagnosis | Post-mortem analysis — what killed specific companies and the warning signs that preceded collapse |
| Thursday | Inside the Machine (AI-Ops) | How Manthan’s AI actually works — the models, the knowledge graph, the reasoning chains |
| Friday | Signal Detection | Early signals the system is flagging right now — companies, sectors, or patterns worth watching |
Intelligence That Compounds
Every Charaka Note connects to every other one through Manthan’s knowledge graph (KG).
The KG isn’t a database. It’s a living network of relationships — companies connected to investors connected to sectors connected to outcomes connected to timing patterns. When we publish a note about competition killing startups, the KG already knows which sectors are most exposed, which companies are in the kill zone, and which investors have seen this pattern before.
This means the notes get more valuable over time, not less. Note 50 will reference patterns from Note 3. Note 100 will surface connections that were invisible at Note 10.
Subscribe early. The compounding starts now.
Who Should Read This
Founders who want to know what actually kills companies like theirs — not what Twitter thinks kills them, but what the data shows.
Investors who want pattern-matched intelligence instead of gut feel. The KG tracks 1,747 investors and their outcome correlations. Some of what it finds will be uncomfortable.
AI builders who want to see a real-world knowledge graph in production — not a demo, not a toy, but a system that cross-references 13,499 entities in real time.
The Promise
“We will never publish an insight we can’t trace back to data. Every pattern comes with its sample size, its confidence interval, and its limitations. If the machine isn’t sure, we’ll tell you it isn’t sure. This isn’t prediction theatre — it’s diagnostic intelligence.”
Welcome to Charaka Notes. The machine is watching. Now you can see what it sees.