Aakash GuptaAI Product Leadership Masterclass with the author of The Making of a Manager
CHAPTERS
Why AI threatens traditional PM/Design roles—and what to do about it
Aakash opens with the existential question: will product designers and product managers still exist in 10 years? Julie reframes the fear as a call to evolve how we think about careers and value creation in product building.
- •AI raises uncertainty about long-term viability of traditional product roles
- •The core question becomes: what steps should practitioners take now?
- •The conversation will focus on becoming an effective AI-era product/design leader
- •Julie’s background (Facebook design leader, author, AI founder) sets context for the discussion
“The death of product development”: from specialized pods to tiny builder teams
Julie explains why AI changes not only products but how teams build them. The classic cross-functional “pod” model (PM, design, eng, research, data) may compress as AI enables individuals to do more end-to-end work.
- •AI accelerates product building capacity per person (automation + ideation)
- •Mature-company model of specialized roles may become less necessary
- •Career ladders and role boundaries will blur as end-to-end building becomes easier
- •Teams may shift toward 1–3 person squads shipping full products
Stop identifying with titles: think in skills, taste, and “product builders”
Julie argues people should detach identity from job titles and instead focus on transferable skills and strengths. The future belongs to “builders” who know what they’re uniquely good at and can orchestrate tools and collaborators to fill gaps.
- •Replace role-identity (“PM/designer/engineer”) with builder-identity
- •Break work into skills and archetypes (systems design, zero-to-one vision, etc.)
- •Identify where you outperform AI today (often taste/judgment)
- •Augment weaknesses via AI tools or complementary teammates
Learning the craft across disciplines to sharpen product taste
They discuss why taste/sense is the differentiator in an AI-assisted world. To direct AI toward exceptional output, leaders must be able to recognize great work and understand what “good” looks like in other functions.
- •The key meta-skill is discernment: exceptional vs average output
- •If you can’t evaluate quality, you’ll default to AI’s average
- •Shadowing and cross-functional learning improves collaboration and judgment
- •Taste becomes more valuable as execution becomes cheaper
A practical system to build taste: find the best, study their work, get critiques
Julie shares a concrete approach: identify world-class practitioners, immerse in their thinking, and validate your mental models through direct feedback. This method applies to any domain (design, analytics, marketing).
- •Start by defining what “amazing” looks like and who does it best
- •Use network recursion: ask who they admire, repeat to find convergence
- •Study outputs and mental models (blogs, talks, artifacts)
- •Reach out for critique to refine your frameworks and standards
What a manager’s job really is: outcomes, not meetings—powered by 3 levers
Transitioning to timeless management principles, Julie redefines management as improving team outcomes toward a goal. She introduces the three durable levers: people, process, and purpose.
- •Management is “getting better outcomes from a group of people”
- •Lever 1: People (hiring, expectations, performance, fit)
- •Lever 2: Process (decision-making, norms, collaboration mechanics)
- •Lever 3: Purpose (vision, alignment, shared definition of success)
AI agents as ‘workforce’: applying people/process/purpose to LLMs
Aakash asks how AI changes those levers; Julie maps agents to management concepts. Selecting models, defining outcomes, and structuring work resembles managing early-career employees—LLMs as “brilliant interns.”
- •Agents can be treated like additional “people” capacity with different strengths/costs
- •Purpose must be clear and narrow: define outcomes for agents explicitly
- •Process matters: structured prompting and task decomposition drive performance
- •Model choice is akin to staffing: different tools excel at different jobs
Calibration questions that prevent misalignment: ‘harder than expected’ vs ‘easier than expected’
Julie explains why expectation alignment is central to human dynamics and effective management. These questions surface mismatched mental models early so leaders can adjust communication, role design, and support.
- •Misalignment often comes from invisible expectations and different worldviews
- •Leaders should test whether their “job preview” matches reality for others
- •Use regular check-ins to expose gaps between perception and reality
- •Better calibration reduces churn, frustration, and downstream conflict
IC → manager blindspots: letting go of doing and thinking in systems
Julie describes the hardest shift for new managers: relinquishing the pride and comfort of IC work. Great managers stop patching isolated problems and instead fix the system that creates them.
- •New managers often keep “fixing” work themselves (doesn’t scale)
- •Shift from local optimization to systemic diagnosis (why does this keep happening?)
- •Use management levers: coaching, accountability, hiring, process improvements
- •Recognize that solutions don’t have to come solely from the manager
Trust and psychological safety: confronting reality without punishing messengers
They explore how leaders build trust by creating an environment where bad news can surface. Julie emphasizes emotional steadiness, gratitude to truth-tellers, and a bias toward action and systemic fixes.
- •Trust grows when leaders can face reality calmly and consistently
- •Avoid taking frustration out on the messenger; reward candor instead
- •Thank people for surfacing problems and collaborate on solutions
- •Balance conviction in direction with openness to uncomfortable truths
Feedback that changes behavior: a gift mindset + a simple delivery script
Julie reframes feedback as holding up a mirror to help someone become who they want to be. The most important factor is genuine care; she also provides a practical structure for delivering feedback clearly and respectfully.
- •Feedback’s goal: help the other person grow into someone they’re proud of
- •Mindset matters more than tactics (not about being right or venting)
- •Delivery script: facts → feelings → assumptions → collaborative resolution
- •Care and respect show up in tone, language, and body cues
Leading through AI disruption: sturdiness, new narratives, and experimentation
Julie outlines what AI-era leadership requires: acknowledging uncertainty, creating a motivating narrative, and treating org design as iterative experimentation. Leaders must also surface and update outdated mental models.
- •AI creates real uncertainty; leaders need flexible ‘willow tree’ sturdiness
- •Craft a narrative that anchors on timeless principles while embracing change
- •Run experiments (e.g., smaller teams shipping end-to-end) and iterate
- •Make mental models explicit—hackathons reveal hidden assumptions about ownership
What great AI product leaders do differently: learn fast, disrupt habits, stay humble
Julie describes the differentiators: strong fundamentals (problem/customer), relentless learning at the frontier, and willingness to be a beginner again. Leaders must model tool adoption and re-earn excellence in the new era.
- •Fundamentals still matter: mission, customers, problem clarity
- •Great leaders actively use AI tools and change their workflows first
- •Experimentation and adaptability become core leadership skills
- •Humility: treat AI era as a reset; rebuild mastery through practice
Essential AI tools and how to think about adopting them (workflows over apps)
Julie lists tools she uses and emphasizes that the bigger unlock is when and how you use them throughout daily work. She advocates frequent experimentation to learn each tool’s strengths and fit.
- •Default: integrate ChatGPT/LLMs for critique, blindspot checks, and drafting
- •Build/prototype tools: Cursor, Lovable, v0; also Claude Code experimentation
- •Meeting capture: Granola; personal life insights: Limitless Pendant (conversation analysis)
- •Adoption strategy: try multiple tools like interviewing candidates; learn strengths/weaknesses
Data, observability, and ‘taste’: how AI changes analytics and what OpenAI does well
Julie explains that data work is about understanding reality with high fidelity—especially when growth is rapid. She shares what she observes working with OpenAI: deep daily interrogation of metrics, rigorous weekly reviews, and strong accountability.
- •Analytics = observability for the business; instrument and model reality early
- •Great teams obsess over “why” behind changes, not just top-line growth
- •Metrics reviews create alignment, accountability, and shared language
- •OpenAI’s approach: high rigor, frequent analysis, and strong operational hygiene
When AI surpasses your taste: the chess analogy and continuing to find meaning
Julie predicts AI will eventually exceed human taste and discusses how to respond. Even if AI is better, humans will still value the learning journey, pride in craft, and the joy of doing—like chess after computers surpassed champions.
- •Today AI is often “better than average” but not yet consistently world-class
- •Use AI more where you’re weaker; train your eye where you’re stronger
- •Assume AI will eventually beat you on taste; prepare by learning from the best
- •Meaning persists in human practice and growth even when AI is superior
Julie’s origin story as a creator: writing as a practice and a ‘letter to self’
Closing out, Julie shares how writing helped her clarify thoughts and communicate more effectively. The book emerged as a way to codify values and learn management through practice, not perfection.
- •Writing compensated for difficulty speaking up early in her career
- •The book was a therapeutic, clarifying “letter to myself” about management values
- •Management/product/design are ongoing practices—execution is harder than theory
- •Continuous improvement comes from showing up, iterating, and learning daily
