Compressing the Talent Stack

AI tools blur role boundaries, enabling individuals to skip communication handoffs and do more

Alexander Embiricos
How to drive word of mouth | Nilan Peiris (CPO of Wise)

Compressing the Talent Stack

"Scott Belsky talks about this idea of compressing the talent stack... maybe the boundaries between these roles are a little bit less needed than before because people can just do much more. And every time someone can do more, you can skip one communication boundary and make the team that much more efficient." - Alexander Embiricos

What It Is

The "Compressing the Talent Stack" framework describes how AI-powered tools are collapsing traditional role boundaries in organizations. When individuals can accomplish more through AI assistance, the handoffs between specialists become less necessary, reducing communication overhead and accelerating execution.

This concept, originally articulated by Scott Belsky and expanded by Alexander Embiricos at OpenAI, suggests that the traditional division of labor between PMs, designers, and engineers is becoming more fluid. A PM can now prototype faster than writing specs. A designer can vibe-code their own animations. An engineer can handle product decisions in the moment rather than waiting for alignment meetings.

The framework challenges the assumption that specialization and clear role boundaries always increase efficiency. In an AI-augmented world, the communication cost between specialists may exceed the benefit of their specialized knowledge.

How It Works

The compression happens in several ways:

  1. Vertical Expansion: Individuals can now do tasks traditionally above or below their role. PMs can code prototypes; engineers can make product decisions; designers can ship features.

  2. Horizontal Expansion: People can operate across adjacent functions. A designer might handle both the design and the frontend implementation. A PM might do both strategy and data analysis.

  3. Eliminated Handoffs: Each handoff between roles carries communication cost, potential misunderstanding, and delay. When one person can do both sides of a handoff, these costs disappear.

  4. Throwaway Work Becomes Viable: Tasks that were "too annoying" for specialists to do (like building a custom data viewer or animation editor) become worthwhile when AI dramatically reduces the effort required.

How to Apply It

  1. Identify your communication bottlenecks - Map where work gets stuck waiting for handoffs between roles

  2. Experiment with role expansion - Give team members AI tools and encourage them to try tasks outside their traditional scope

  3. Measure cycle time, not utilization - Focus on how fast work flows through the system, not how busy each specialist is

  4. Prototype before speccing - When building something new, have the closest person to the problem build a rough version with AI rather than writing requirements for someone else

  5. Create "throwaway work" culture - Encourage building quick custom tools (data viewers, animation editors, analysis scripts) that would have been too expensive before

When to Use It

  • When forming small, fast-moving teams
  • When deciding team structure for AI-era companies
  • When individual contributor velocity matters more than coordination
  • When you notice work frequently getting stuck in handoff queues
  • When evaluating how AI tools should change your org design

Source

  • Guest: Alexander Embiricos
  • Episode: "How to drive word of mouth | Nilan Peiris (CPO of Wise)"
  • Key Discussion: (00:44:24) - Discussion of how Codex changes PM work and role boundaries
  • YouTube: Watch on YouTube

Related Frameworks

Back to Team Building