Nikita Shamgunov, former leader of Neon who pivoted the company to AI-driven Postgres and later joined Databricks. He tells the story of Replit agents spinning up databases far faster than humans and how that forced a rapid company pivot. The conversation covers rebuilding search and infrastructure for AI agents, rapid AI hiring and execution, and safe models for agent autonomy.
25:54
forum Ask episode
web_stories AI Snips
view_agenda Chapters
auto_awesome Transcript
info_circle Episode notes
insights INSIGHT
Web Needs An Agent Native Index
The web must be rearchitected for AI agents as primary customers rather than humans.
Parag built Parallel because agents work differently than humans and the index must complement models with different crawling and ranking.
insights INSIGHT
Search Indexes Built For Humans Break For Agents
Existing search indexes and ranking are optimized for human behavior and fail for agent workflows.
Parag argues an index should be owned end-to-end and designed as a complement to large models' parametric memory.
question_answer ANECDOTE
Replit Agents Created Databases 4x Faster
Replit Agents caused Neon to see databases created at 4x the human rate, revealing a new primary user: coding agents.
Neon noticed usage spikes were throwaway apps, high platform retention but low per-app retention, highlighting serverless fit.
Get the Snipd Podcast app to discover more snips from this episode
What happens when AI agents — not humans — become your primary customer? That's not a hypothetical. It's already happening, and the founders who recognize it earliest are rebuilding their entire infrastructure stacks from scratch.
In this live episode of Founded & Funded from our IA Summit in Seattle, Madrona Venture Partner Jon Turow sits down with Parag Agrawal, former CEO of Twitter and founder of Parallel Web Systems, and Nikita Shamgunov, who led Neon through a rapid AI pivot before its acquisition by Databricks.
What they cover:
Why Parag is building a new search index from the ground up — and why existing ones weren't designed for AI agents
The moment Nikita realized Replit agents were spinning up databases 4x faster than all human developers combined — and what that forced him to do
How to pivot an established company in weeks, not months, when your customer base suddenly changes
The "pagers vs. iPhones" framework for knowing when to lean into disruption vs. protect what you have
Parag's two-person hiring rubric for teams operating in deep uncertainty
Why Nikita added the head of product for ChatGPT to Neon's board — and what that signaled to the market
The "two-way door" model for giving agents real autonomy without catastrophic downside
Whether you're building infrastructure, running an AI-native startup, or trying to figure out where your product fits in an agent-first world — this conversation will sharpen your thinking.
Full Transcript: https://www.madrona.com/twitter-ex-ceo-web-built-for-humans-make-it-work-for-ai-agents-nikita-Shamgunov-parag-agrawal
Chapters (00:00) – Introduction (01:52) – Parag Agrawal: Why Parallel Was Built for AI Agents From Day One (03:22) – Why Existing Search Indexes Don't Work for AI Agents (05:08) – Nikita Shamgunov: How Replit Agents Outpaced the Entire World on Neon (08:27) – The Pager-to-iPhone Decision: Lean Into Disruption or Get Left Behind (11:13) – How Neon Built an AI Team in Two Weeks and Launched MCP Before Anyone Else (13:41) – Firing Bullets: Why a 4-Out-of-9 Batting Average Was Good Enough (15:37) – Parag on the Two Types of People You Need to Take Concentrated Risk (21:08) – Building Trust in Agents: Evals, Confidence Scores, and Read-Only Infrastructure (23:32) – Nikita's Two-Way Door Framework for Agent Autonomy (25:35) – Parallel Execution: Fork Environments and Let Agents Compete
Founded & Funded