CHAPTERS
- 0:05 – 0:35
Sim.ai in one sentence: a Figma-like canvas to build AI agents
Amir explains Sim.ai as a visual, end-to-end agent builder: a canvas for composing agent workflows with drag-and-drop plus natural-language prompting. The goal is to make building agents usable both inside products and for background automation.
- •Visual canvas for orchestrating agentic workflows (Figma-like UI)
- •Cursor-style side prompt + natural-language building
- •Drag-and-drop workflow composition
- •Use cases: embed agents in-product or automate internal tasks
- 0:35 – 1:00
Current traction: 60,000 developers and fast-growing open source adoption
Amir shares the company’s current metrics and why they focus on developer adoption. He highlights rapid open source growth as a key signal of momentum.
- •60,000+ developers on the cloud platform
- •17.5K+ GitHub stars; among fastest-growing open source workflow builders
- •YC batch was recent, growth has accelerated quickly
- •Metrics they care about: developer usage + open source traction
- 1:00 – 2:21
Founding story: meeting at Berkeley and the leap to start a company
Amir recounts meeting co-founder Waleed through a Berkeley roommate connection and building a close friendship before working together. After graduation, Amir convinces Waleed—then at Amazon—to move to SF and start a company.
- •Met at Berkeley via roommate’s high-school friend connection
- •Became best friends; spent years working/hanging out daily
- •Amir calls Waleed to move to SF; Waleed quits and relocates
- •They hadn’t built projects together before co-founding
- 2:21 – 4:08
First idea (and first YC application): personalized landing pages that didn’t fit
The initial startup concept was dynamically personalized websites based on CRM/browsing data, targeted at sales and marketing teams. They learned quickly they didn’t understand the customer deeply, and their YC interview exposed how confusing the product was to explain.
- •Concept: individualized landing pages tailored per visitor
- •Tried selling into sales/marketing without domain understanding
- •Applied to YC with this idea; interview revealed unclear messaging
- •Post-interview realization: they needed a new direction
- 4:08 – 5:20
The seed of Sim: building Scalemont required “hundreds of agents”
While working on the earlier project, they repeatedly assembled many small agents (research, copy generation, image prompting, etc.). The complexity and lack of shared visibility into agent components became the real pain point—and the inspiration for a visual orchestration layer.
- •Prior product required chained agents for research, copy, images, and sections
- •Workflow complexity became hard even for the two founders to manage
- •Hard to reuse/track: “did you already build an agent for that?”
- •Insight: agents are the core problem worth solving
- 5:20 – 8:35
Deciding to build an AI-native workflow platform (Figma × Jupyter)
After exploring other industries, Amir leans into what they know best: LLMs and agents. He frames Sim as combining the ease of a design canvas with programmable notebook-like power, initially as a five-day experiment that quickly became the company.
- •Pivot lesson: don’t reach too far outside your expertise
- •Vision: merge Figma-like UI with notebook-like building for agents
- •Started as a 5-day test; co-founder initially skeptical
- •Early expectations were tiny (10–100 GitHub stars) compared to reality
- 8:35 – 10:50
Differentiation: moving beyond “traditional automation” to AI-native workflows
Amir contrasts Sim with legacy workflow automation tools, arguing they aren’t designed for agent loops, large-scale repetition, or natural-language construction. He describes workflows as the next abstraction layer above code, following trends like AI-assisted programming.
- •Traditional workflow tools exist but aren’t AI-native in key ways
- •Need: run one agent pattern 100x, then synthesize results
- •Need: build workflows via natural language, not just rigid connectors
- •Belief: workflows become the future programming paradigm as abstraction rises
- 10:50 – 12:42
Open source growth playbook: ethos + distribution where developers are
Sim’s open source strategy centers on being genuinely open (license and intent) and showing up in developer communities. Amir argues code is becoming less proprietary in an AGI/LLM world, so the value shifts to ecosystem, network, and execution.
- •True open source (e.g., Apache/MIT) builds developer trust
- •Thesis: code is less defensible; vision + network matter more
- •Goal: be foundational workflow “engine” others can adapt by industry
- •Distribution: Hacker News, Twitter, Discord; community-first engagement
- 12:42 – 16:02
Engineering launches for GitHub Trending—and converting to cloud revenue
Amir explains how coordinated launches can push a repo onto GitHub Trending, driving the majority of star bursts. He then connects open source attention to cloud-platform signup growth and monetization via hosted convenience and paid inference.
- •GitHub Trending drives large daily star spikes (hundreds to thousands)
- •Strategy: orchestrated, simultaneous pushes (HN, Product Hunt, Twitter)
- •Stars and cloud signups tracked closely early on
- •Monetization path: cloud convenience + paid inference vs. self-hosting
- 16:02 – 21:15
Operating cadence post-YC: ship every two weeks, keep urgency, hire for passion
The team sustains momentum with a strict launch calendar and a culture that favors shipping over perfection. Amir also describes what they look for in hires: curiosity, excitement about AI’s future, and the ability to execute regardless of specific tech stack familiarity.
- •Hard part: staying relentlessly better while avoiding perfection traps
- •Launch calendar: ship something every two weeks (SNL “it’s Saturday night” mindset)
- •Post-YC acceleration: grew from ~5K to ~60K developers in months
- •Hiring: passion, curiosity, shipping mindset; skills can be learned or swapped (Go vs TS)
- 21:15 – 25:05
Founder lessons: think in decades, stay healthy, optimize for long-term learning
Amir reflects on personal changes: embracing long time horizons, building sustainable routines, and prioritizing learning that compounds. He emphasizes choosing a mission you’d keep doing even through low points—and aligning roles with what you’re best at.
- •Great companies take ~10 years; compounding daily effort matters
- •Sustainability: balance, health, gym, food—avoid burnout operating mode
- •Be honest about what you understand/don’t understand; keep learning
- •Choose a path you’d stick with for years; avoid short-term mismatches (sales/marketing detour)
