⚡️ 10x AI Engineers with $1m Salaries — Alex Lieberman & Arman Hezarkhani, Tenex
Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
Podcast Summary
Alex Lieberman and Arman Hezarkani, co-founders of 10X, are revolutionizing software consulting by compensating AI engineers based on output rather than hours worked. (01:29) Their breakthrough came when Arman downsized his previous company's engineering team by 90% but actually increased production-ready software output by 10x through AI-first processes. (02:44) This counterintuitive result revealed that traditional hourly compensation models create perverse incentives that discourage engineers from leveraging AI tools to work faster. (04:04) Now, 10X expects multiple engineers to earn over $1 million annually through their story point-based compensation system while delivering unprecedented speed and quality for clients. (09:18)
- Core theme: Transforming knowledge work compensation models in the AI era to align incentives with productivity gains
Speakers
Alex Lieberman
Alex Lieberman is the co-founder of Morning Brew, a successful business newsletter that became a major inspiration in the newsletter industry. He invested in Arman's previous company Parthian in 2020, which led to their partnership in founding 10X. Alex brings entrepreneurial experience and a non-technical perspective to understanding AI's transformational potential in business.
Arman Hezarkani
Arman Hezarkani is a software engineer and entrepreneur who previously founded Parthian, an AI financial tools business. After being forced to downsize his engineering team by 90%, he discovered that AI-first processes could actually increase software output by 10x. This experience led him to co-found 10X with Alex, where he focuses on the technical implementation of their revolutionary compensation model.
Key Takeaways
Compensation Models Must Align with AI-Enabled Productivity
Traditional hourly billing creates perverse incentives where engineers are discouraged from using AI tools that make them more productive. (04:04) When engineers are paid by the hour, they're economically incentivized to work slower, even when AI could enable 10x faster output. 10X solves this by compensating engineers based on story points (completed output) rather than time spent, creating direct economic incentives to adopt new AI tools and optimize workflows. This approach enables some engineers to earn over $1 million annually while delivering exceptional value to clients.
Quality Control Through Competing Incentives
To prevent gaming of output-based compensation, 10X employs technical strategists who are incentivized based on client retention (NRR) and serve as the final quality gate before any work reaches clients. (07:37) This creates healthy tension between engineers maximizing story points and strategists ensuring long-term client satisfaction. Additionally, they hire for "long-term selfish" engineers who understand that inflating story points will destroy client relationships and their future earning potential.
AI Enables Unprecedented Speed in Sales and Development
AI-powered rapid prototyping fundamentally changes sales motions and client relationships. (11:07) When a fitness influencer initially rejected 10X's services, one engineer built a working prototype of their desired health coach app in just four hours, immediately moving 10X to the top of the prospect's vendor list. This speed advantage allows for "show don't tell" sales approaches that were previously impossible, where working products can be delivered faster than traditional sales cycles.
High Structure Enables Longer Autonomous Agent Work
TypeScript's combination of JavaScript flexibility with type constraints allows AI coding agents to iterate autonomously for longer periods. (12:42) The error messages and type checking provide feedback loops that enable agents to self-correct and continue working without human intervention. This architectural choice significantly amplifies the productivity gains from AI tools and reduces the need for constant human oversight in the development process.
Entropy Control Is the Key to Autonomous AI Systems
The most sophisticated challenge in building truly autonomous AI engineers isn't model intelligence or context length, but controlling entropy - the accumulating error rate that derails agents over time. (20:18) Even a 1% error rate compounds and multiplies, eventually causing complete system failure. The solution requires achieving near-100% accuracy for individual tasks to prevent error accumulation, which represents a fundamental engineering challenge beyond simply having smarter models.
Statistics & Facts
- 10X downsized their previous company's engineering team by 90% but increased production-ready software output by 10x through AI-first processes. (02:44) This counterintuitive result demonstrated the massive leverage potential of properly implemented AI tools in software development.
- Multiple 10X engineers are expected to earn over $1 million in cash compensation next year purely from story point earnings. (09:18) This showcases the significant financial upside possible when compensation aligns with AI-enabled productivity gains.
- Snapback Sports' mobile trivia app built by 10X hit 20th place globally on the App Store after being developed in just one month. (10:47) This demonstrates the speed and quality achievable through their AI-enhanced development processes.
