Paying Engineers like Salespeople – Arman Hezarkhani, Tenex

AI Transformation & Engineer Compensation at 10X

Key Concepts:

• Story Points: A unit of measure for expressing the overall effort required to implement a particular feature or task in software development.
• AI Transformation: The process of integrating Artificial Intelligence technologies into a company’s operations, products, and services.
• Incentive Structures: Systems designed to motivate individuals or teams to achieve specific goals.
• Quantization (of models): Reducing the precision of the numbers used to represent a machine learning model, making it smaller and faster to run.
• NR (Net Revenue): A metric used to measure customer satisfaction and the value delivered by a service.

The Shift in AI Compensation: From Hourly to Output-Based

Arman, co-founder and managing partner at 10X, discusses a novel approach to compensating engineers – tying pay to the completion of story points, mirroring a sales commission structure. This model arose from observing a stark contrast in efficiency between engineers: some leveraging AI tools extensively, while others struggled with basic coding tasks. He argues that traditional compensation models (hourly, salary + bonus, equity) fail to adequately incentivize the adoption and effective use of AI in software development.

He highlights a personal experience at a WeWork in New York, witnessing a fellow engineer painstakingly typing code character by character while he himself utilized 45 AI agents for tasks like lunch ordering, code writing, and research. This disparity underscored the need for a system that rewards maximizing output through AI integration.

A Historical Perspective on Software Engineer Compensation

Arman outlines a brief, illustrative history of software engineer compensation:

• Hourly Pay: The initial model, deemed broken due to its lack of incentive for faster output and potential for inflated time estimates. Engineers had a disincentive to complete work quickly.
• Salary + Bonus: Led to a “punch-in, punch-out” mentality, lacking the drive for exceptional performance.
• Equity: Effective for incentivizing innovation, particularly in startups, but reliant on successful exits and not universally applicable. Many engineers prioritize immediate cash compensation over the potential, but uncertain, rewards of equity.

He contends that the current model is outdated in the age of AI and requires reinvention to directly incentivize AI tool adoption and maintain high code quality.

The 10X Compensation Model: Strategists, Engineers & Story Points

10X operates with two primary client-facing roles: Strategists and AI Engineers.

• Strategists: Former PMs and engineers who translate client product requirements into actionable plans. They focus on product management and consulting.
• AI Engineers: Responsible for the technical implementation, starting with an architecture design document.

Work is broken down into tickets, each assigned a value in story points – a standard metric representing the effort required. Engineers receive a base salary plus quarterly bonuses based on completed story points.

The process unfolds as follows:

1. Client Request: Clients initiate projects with broad requests (e.g., “We want AI”).
2. Roadmapping: 10X develops a roadmap outlining potential AI solutions.
3. Execution: Based on the roadmap, work is broken down into tickets with assigned story points.
4. Implementation: Engineers build and implement the solutions.
5. Compensation: Engineers are paid a fee per completed story point, in addition to their base salary.
6. QA & Approval: Every ticket requires multiple rounds of Quality Assurance (QA) and client approval, with Strategists involved in the review process.

Case Studies: Real-World Applications & Results

Arman presents two case studies demonstrating the effectiveness of the model:

• Billboard Company: 10X developed an AI model for automated image moderation, reducing the need for human reviewers and accelerating revenue generation. The model achieved 96% accuracy compared to human moderators and was delivered in two weeks.
• Retail Analytics Company: 10X built five AI models that run on low-power devices in retail stores, providing insights like heat mapping, queue detection, and theft detection. This expanded the functionality of existing devices and opened new revenue streams.

These projects highlight the speed and quality of delivery achievable with the story point-based compensation system.

Potential Risks & Mitigation Strategies

Arman acknowledges potential risks associated with the model:

• Story Point Inflation: Engineers potentially overestimating story points to increase earnings. Mitigation: Strategists scope the work and review estimates.
• Reduced Code Quality: Engineers rushing to complete story points, compromising quality. Mitigation: Rigorous QA processes, client approval, and the involvement of Strategists.
• Competitive/“Sharp Elbowed” Behavior: Engineers competing aggressively for story points. Mitigation: Careful hiring practices and a focus on building a collaborative team.

He emphasizes the importance of hiring the right people – individuals who are motivated by AI and committed to quality – as the foundation for success. He quotes his co-founder, Alex, stating, “AI makes people look like one of those crazy mirrors where any one of your attributes, it makes it 10 times larger. If you're a great engineer, AI makes you great. If you're not, it makes you sloppier.”

Conclusion: Unlocking Potential in the Age of AI

Arman concludes that the current compensation models may be hindering employees from fully leveraging the potential of AI. He advocates for exploring alternative incentive structures that directly reward the effective use of AI tools while maintaining high quality standards. He believes AI provides “superpowers” and encourages listeners to consider how they can unlock their team’s potential through innovative compensation strategies. He offers his contact information (arman@10x.co) for further discussion.