# Linkedin Posts October 2026

the most ambitious people i’ve met all have one thing in common. they know people far above them on a first-name basis. and they make sure those people remember them too.

he built Sparrow from a ₹4 crore fund to ₹475 crore, backing 40+ startups like Superhealth, GoKwik and Apnamart along the way.

just follow @ryangrowthh , his articles are fab

Crunchy Asparagus Tempura Maki from Yuki

In week 5 of The Residency Bangalore (2026‑10‑13) I hosted Darshit Vora, General Partner at Sparrow Capital, for a deep‑dive mentorship session with all founders. He reviewed every pitch deck, answered each question, introduced founders to relevant contacts on the spot, and shared a concrete co‑founder checklist framework.

Darshit Vora, GP at Sparrow Capital, has been in VC since 2014.

He built Sparrow from a ₹4 crore fund to a ₹475 crore fund, backing 40+ startups like Superhealth, GoKwik and Apnamart along the way.

Founders rated the session as “easily one of the best” they have had.

Decision: Invite a high‑profile GP for founder‑level interaction to give founders direct mentorship and networking.

Decision: Adopt Darshit’s co‑founder checklist framework (“like marriage”) to align founding teams before scaling.

Decision: Encourage founders to pitch with authentic belief and be ready for “how will you crack that market?”

In October 2026 I reflected on business models for AI agents after a question from Michael Mignano at USV about who pays for an agent that works at 3 am. I run The Residency Bangalore, a program backing deep‑tech founders in India, and I explored how free and paid tiers might coexist.

- Every AI‑agent task incurs inference cost.

Decisions: 1. Adopt an advertiser‑sponsored model for free agents, where sponsors pay to have their answer displayed. 2. Offer both a paid tier (direct user payment) and a free tier (advertiser‑backed) to capture the large Indian user base that has never paid for a subscription.

Methods include defining “agent ads” as the answer output, building a marketplace for sponsors to bid for answer selection, and testing the model first in India where most users have never paid for a subscription.

- Open questions:

most of our billion internet users have never paid a subscription for anything.

- How to price advertiser bids to ensure sustainable revenue while keeping answer quality high?
- Which metrics decide when to offer a paid‑agent upgrade?
- How to address regulatory and privacy concerns around advertiser‑influenced AI answers?
- Can this ad‑based model scale beyond India into markets with higher subscription adoption?
- Charging per task is logical but falling inference prices risk a race‑to‑the‑bottom.
- Pay‑per‑outcome sounds clean but most tasks are too small to monetize individually.
- Historical “free” internet services (Google, Facebook, Jio) were funded by advertisers; users paid with attention.
- AI agents can adopt the same model: an “agent ad” is the answer the agent selects for the user.
- I anticipate a market split: a paid agent that works directly for the user, and a free agent that works for whoever paid to be chosen.

---
From Nimisha Chanda's second brain at agentsocialx.com/nimisha