Flo Crivello's AI Agent Philosophy
Table of content
Flo Crivello builds AI assistants that actually do work. Not chatbots that answer questions—agents that handle your email, schedule meetings, process documents, and run sales outreach while you sleep.
As founder and CEO of Lindy.ai, he’s betting that the future of personal AI isn’t about having conversations with machines. It’s about delegating real tasks to them.
From Uber PM to AI Founder
Crivello joined Uber in 2015 as a product manager. He later founded Teamflow, a virtual office startup, before pivoting to AI agents with Lindy.
The path wasn’t linear. His blog chronicles the messy reality of building startups: the pivot from virtual offices to AI automation, the painful decision to move his team back to San Francisco after going fully remote, and his contrarian stance that co-founders might actually hurt more than help.
What Makes Lindy Different
Lindy isn’t another LLM wrapper. It’s an agent builder where you describe what you want done, and the system creates an AI employee to handle it:
- Support agents that resolve tickets without human intervention
- Sales agents that research, qualify, and reach out to leads
- Document processing that turns paperwork into structured data
- Voice agents for phone automation
The bet: most business workflows are automatable once you give AI the right tools and permissions. Lindy provides 3000+ integrations and a no-code builder to wire them together.
The Solo Founder Case
Crivello makes a provocative argument against co-founders in his essay “Co-Founding Considered Harmful”:
“Solo founders are twice as likely to be successful than co-founders… almost all co-founders I know have broken up, often endangering the company.”
His reasoning: the co-founder search space is too limited, the stakes too high, and separation too painful. Instead of finding “another you,” he suggests over-incentivizing early hires with 2-4% equity to get the same buy-in without the legal complexity.
On AGI Risk
Unlike many AI builders who dismiss safety concerns, Crivello takes a worried stance:
“The folks closest to the problem are the least able to worry… caught in a testosterone-fueled race to the biggest models.”
He sees a troubling pattern: humans can’t reason about exponentials, AI researchers are too close to feel the danger emotionally, and the broader tech community is stuck in “mood affiliation”—wanting to be optimists and reasoning backward from that feeling.
Practical Takeaways
1. Build for delegation, not conversation
The real value of AI isn’t answering questions—it’s taking tasks off your plate entirely. When designing AI workflows, ask: “Can this run without me checking in?”
2. Network effects still matter
Crivello argues for San Francisco not because of the startup mythology, but for network effects. Being at the center of your industry compounds: better information, better hires, better opportunities. “It’s unclear why one would decide to be anywhere else.”
3. Create stress before committing
If you’re considering a business partner, put yourselves in artificially stressful situations first. “Try too little food, too little sleep, too little space—maybe extreme camping. People reveal their true selves under high stress.”
Resources
- Lindy.ai - Build AI agents without code
- Flocrivello.com - Blog on startups, AI, and philosophy
- @Altimor on Twitter - Regular takes on AI and building
- Co-Founding Considered Harmful - The case for going solo
- Go West, Young Man - Why location still matters
Get updates
New guides, workflows, and AI patterns. No spam.
Thank you! You're on the list.