A Conversation with Tomer Cohen, Former Chief Product Officer, LinkedIn

Key Concepts

• Full-Stack Builder: An organizational archetype where individuals possess cross-functional skills (coding, design, product management) to execute tasks end-to-end, replacing siloed roles.
• AI Agency: The proactive use of AI to automate workflows and solve problems, moving beyond mere "fluency" (talking about AI) to actual implementation.
• Bottom-Funnel Metrics: Shifting focus from input metrics (token usage, adoption rates) to outcome-based metrics (revenue growth, engagement, productivity gains).
• AI Traffic Control: A governance layer within companies that determines which AI models are appropriate for specific tasks to optimize cost and performance.
• Skill Gap: The widening disparity between the skills required by the rapidly evolving labor market and the current capabilities of the workforce.

1. The Shift in Organizational Structure

Tomer Cohen argues that traditional corporate structures, characterized by rigid functional silos (PM, Engineering, Design), are becoming obsolete.

• Collapsing the Stack: Large companies suffer from "process complexity" where simple tasks require multiple reviews and handoffs. By adopting a "Full-Stack Builder" model, organizations can reduce this complexity.
• Navy SEAL Analogy: Instead of large, specialized teams, companies should aim for small, mission-oriented "pods" of builders who can tackle emerging priorities without waiting for cross-departmental approvals.
• System Builders vs. Specialists: While the need for generalist "Full-Stack Builders" is rising, there remains a role for "System Builders" (infrastructure) and a smaller, elite group of "Specialists" who function like symphony musicians—needed only for high-level, complex tasks.

2. AI Agency and the "Mid-Career" Risk

Cohen presents a provocative perspective on who is most at risk in the AI era:

• Early-Career Talent: Often "AI-native," highly malleable, and unburdened by legacy best practices. They are currently the most adaptable workforce segment.
• Mid-Career Talent: Identified as the most vulnerable group. They have built their careers on "best practices" that are now being disrupted, and they often exhibit an "adversity to change."
• The "Vibe Coding" Hackathon: To force internalization of AI, Cohen implemented mandatory 15-hour hackathons for leadership. He noted that even highly technical leaders did not truly grasp the power of AI until they "felt it in their hands."

3. Measuring AI Value: From Consumption to Outcomes

Companies are moving past the "AI Theater" phase (where usage is encouraged for the sake of it) into a phase of rigorous ROI evaluation.

• The Funnel Approach:
  • Top: Input metrics (tokens, adoption).
  • Middle: Operational metrics (number of experiments, PRs completed).
  • Bottom: Business outcomes (revenue growth, engagement).
• The Compute-People Cost: CFOs must now manage a cost structure composed of both human capital and compute. Companies like Meta and Google have mastered this by linking AI compute directly to ad revenue, whereas SaaS companies are struggling to find a clear "outcome-based" pricing model to replace seat-based licensing.

4. Strategic Advice for MBAs and Entrepreneurs

• Where to Bet: For those entering the workforce, Cohen advises ignoring titles and compensation. Instead, prioritize companies that are "uncomfortable" and pushing the envelope. If a company is not actively re-engineering its workflows, it is an "old-world" entity.
• The "SAS Apocalypse": While some argue enterprise SaaS is dead, Cohen suggests that if a company delivers measurable productivity gains and possesses unique data, it remains viable. The "moat" is no longer the AI model itself (which is becoming a commodity), but the unique governance, ethical boundaries, or domain-specific workflows built around it (e.g., the legal-tech firm Harvey).
• Developing Judgment: Judgment is built through experience, failure, and observing mentors. In an era where "grunt work" is automated, junior employees must find new ways to gain intuition, such as running experiments and actively seeking out high-pressure environments.

5. Notable Quotes

• "Change is happening much faster than we’re able to respond to it."
• "Using ChatGPT is not agency. My daughter is seven years old; she uses ChatGPT. That’s not agency."
• "If you’re joining a company where they’re not thinking about this [AI], you’re going to be building for old problems with old techniques."
• "Good ideas will start becoming bottlenecks because there’s so much you can push, but are you pushing the right things?"

Synthesis

The core takeaway is that the AI revolution is not merely a technological upgrade but a fundamental shift in how work is organized and valued. Success in this era requires a "beginner’s mindset," the ability to move from "AI fluency" to "AI agency," and a willingness to abandon legacy best practices in favor of agile, outcome-focused building. Organizations that fail to collapse their functional silos and move toward outcome-based metrics will likely struggle to remain competitive as the "skills gap" continues to widen.