Episode Summary
Executive Summary: Tomer Cohen explains LinkedIn’s “Full Stack Builder” model: a new way of building product that uses AI to collapse process and organizational complexity, empower smaller cross-functional pods, and automate everything except human judgment, empathy, vision, communication, and creativity. The conversation covers platform re-architecture, custom agents, culture change, performance incentives, and why companies must rethink product development now.
Main Topics: Why LinkedIn is rethinking product building (Priority: 5/5): Tomer argues that the pace of change is outpacing organizations’ ability to respond, making traditional product-development models too slow and too specialized for the AI era. The Full Stack Builder model (Priority: 5/5): LinkedIn’s new operating model lets builders take an idea from insight to launch across disciplines, using AI and smaller pods rather than large siloed teams. What humans should do vs. what AI should automate (Priority: 5/5): The model prioritizes human strengths like vision, empathy, communication, creativity, and especially judgment, while automating most of the execution and coordination work around them. Platform, tools, and agents (Priority: 4/5): Success depends on re-architecting internal systems, building custom agents (trust, growth, research, analyst, maintenance), and integrating them tightly with LinkedIn’s context rather than relying on off-the-shelf tools. Culture and change management (Priority: 5/5): Tomer stresses that adoption requires incentives, examples, training, performance expectations, and visible wins—tools alone will not change behavior. Pilot results and organizational rollout (Priority: 4/5): The program is already saving hours, improving quality, and attracting top talent; LinkedIn is formalizing a career path, pods, and an associate program to scale it. Career implications and future of specialization (Priority: 4/5): LinkedIn is creating a full stack builder title and reducing reliance on traditional specialization, while still preserving room for specialists who prefer that path.
Key Arguments: By 2030, job skills will shift so much that even people staying in the same role will need to relearn large parts of their job. Product work has become overcomplicated through process complexity and organizational micro-specialization, not because the underlying work is inherently complex. AI should not replace the core human traits of building; it should automate everything except high-quality judgment, empathy, vision, communication, and creativity. Off-the-shelf AI tools are insufficient for a large, legacy company like LinkedIn; they must be customized to internal code, design systems, and knowledge bases. The most effective AI adoption strategy is not just tool deployment but change management: incentives, performance reviews, training, visible examples, and cultural reinforcement. Small cross-functional pods with full-stack builders can move faster, adapt better, and produce higher-quality work than traditional large functional teams. Top talent is often the first to adopt and benefit from AI tools because they are motivated to keep improving their craft. The future is not everyone becoming a full stack builder; rather, more people should have enough fluency to flex across roles when needed, while some specialization remains. Large-scale transformation requires patience and upfront investment in platform, customization, and internal enablement before productivity gains show up.
Data Points: Skills change by 2030: 70% - Tomer says the skills required for a job will change by 70% by 2030. Fastest-growing jobs growth: north of 70% - He says the most in-demand jobs are growing by more than 70% year over year from last year’s fastest-growing jobs. Impact on jobs: 90–95% - He notes some jobs will be far more affected by AI than others, while roles like nurses may see less impact. Sources used for research: 10 to 15 - LinkedIn product teams often use 10–15 sources of information before feeling they’ve researched a problem well. Time to first MVPs: 4 to 5 months - He says the first MVPs of the new agents arrived about four to five months after focused work began. Build automation: close to 50% - Maintenance agent and QA agent are now handling close to half of all builds. Organization adoption: substantial part of the org - He says a substantial part of LinkedIn is already using the tools, though not yet a high percentage company-wide. Time savings: hours per week - Early adopters like PMs, designers, and engineers are saving hours of work each week. Program rollout: next couple of months - He expects the internal program to GA in the next couple of months. APM replacement: starting January - LinkedIn’s APM program ends this year and the associate product builder program begins in January.
Pivotal Quotes: "the time constant of change is far greater than the time constant of response" — Tomer Cohen: Explains why LinkedIn believes current operating models are too slow for the pace of technological change. "I want to automate everything outside of those five traits that we talked about" — Tomer Cohen: Describes the division of labor between humans and AI in the full stack builder model. "if you're waiting for a reorg, you're not thinking about it the right way" — Tomer Cohen: His advice to employees and leaders who want to start building differently now rather than wait for formal structure changes.
Implications: The episode suggests AI transformation will be won by companies that redesign workflows, incentives, and team structure—not just buy tools. Builders who develop AI fluency and judgment will gain the most advantage.
About Lenny's Podcast
Lenny Rachitsky interviews world-class product leaders and growth experts about building products and growing careers.