Invest Like the Best with Patrick O'Shaughnessy

Gokul Rajaram - Lessons from Investing in 700 Companies - [Invest Like the Best, EP.456]

January 29, 2026

Key Takeaways Copied to clipboard!

  • Product development is shifting to a bottoms-up, hands-on approach where Product Managers must understand and prototype with AI capabilities, with judgment being the most future-proof human skill. 
  • Legacy software companies priced on utility (like Zendesk) are highly exposed to AI disruption, whereas those tied to timeless data or financial flows (like ERPs) are more insulated. 
  • Durable business models in the AI era rely on owning scarce assets, control points, network effects, or integrating financial services, as building on top of large platforms risks being squeezed. 
  • A North Star Metric must balance customer value and business value, and should always be coupled with 'check metrics' (like margin or retention) to prevent unintended negative consequences from optimization. 
  • The best software companies must be fully self-serve, as opening systems to self-service users leads to faster product improvement through exposure to sophisticated, unexpected use cases. 
  • In the AI era, career success favors functional experts who focus on 'doing and building' and learn to orchestrate armies of AI agents, rather than hiring middle management with small spans of control. 

Segments

AI Impact on Product Development
Copied to clipboard!
(00:06:11)
  • Key Takeaway: Product development is shifting from rigid roles to a hands-on, bottoms-up approach where PMs must prototype and own evaluation (evals) of non-deterministic AI outputs.
  • Summary: The traditional separation of PM, design, and engineering is blurring, with PMs now checking in code and design roles potentially shrinking as AI leverages design systems. Non-deterministic software requires PMs and researchers to own ’evals’ to judge the reasonableness of AI-produced outputs across various use cases. Capabilities are advancing so fast that rigid planning fails; constant hands-on iteration is necessary.
Product Philosophy and Outcomes
Copied to clipboard!
(00:11:12)
  • Key Takeaway: The core job of a product person is balancing customer needs and business needs, defined by measurable customer behavior changes that serve as leading indicators for business outcomes.
  • Summary: Product management requires balancing customer value against company value, avoiding building things that are only good for one side. The best product people focus on outcomes defined by customer behavior changes, such as moving a user from a non-customer state to a loyal state. Every feature launch must be grounded in a clear hypothesis articulated as a specific, measurable customer behavior shift.
Future-Proofing in AI Era
Copied to clipboard!
(00:13:55)
  • Key Takeaway: Judgment is the single most future-proof human skill, essential for evaluating the massive volume of code and output generated by AI engineers to prevent ‘AI slop.’
  • Summary: The primary challenge in the age of infinite productivity is determining what matters and ensuring the right things are built, requiring human judgment in code review and design evaluation. Engineers must still review AI-generated code for correctness and vulnerability, and PMs must apply judgment to ensure outputs align with the broader design system and user needs.
Building Durable AI Applications
Copied to clipboard!
(00:16:44)
  • Key Takeaway: Durable AI applications must target deep, complex workflows requiring custom data and secure ownership of a scarce asset or control point to avoid being absorbed by horizontal foundation models.
  • Summary: Successful AI applications must solve high-value, complex problems that foundational model builders cannot easily replicate using generic agent tools. Durability requires owning a scarce asset, controlling a financial or data flow, possessing unique hardware, or building an entire system of record rather than just workflows on top of incumbents. Agent companies built on top of systems of record (like Slack) risk being cut off or commoditized.
Software Stickiness in AI Age
Copied to clipboard!
(00:27:25)
  • Key Takeaway: Product stickiness in the AI era is derived from network effects, financial flows, hardware integration, or access to unique, non-replicable assets.
  • Summary: Network effects, like those in DoorDash, create durability because an AI cannot easily replicate the entire ecosystem of users, restaurants, and dashers simultaneously. Products that handle money movement, such as business banks or POS systems like Toast (which bundles hardware), create high switching costs. Stickiness is also achieved by controlling access to unique assets, such as a highly influential individual like Brett Taylor at Sierra.
Lessons from Tech Leaders
Copied to clipboard!
(00:29:41)
  • Key Takeaway: Generational company leaders possess a superpower perfectly aligned with their company’s core needs, exemplified by Larry and Sergey’s focus on technological superiority and Mark Zuckerberg’s insight into consumer engagement.
  • Summary: Larry Page and Sergey Brin prioritized technological superiority and massive scale (e.g., 1GB Gmail storage) over immediate revenue, investing for the long term. Mark Zuckerberg excels at critiquing consumer product flows for engagement and derived the foundational ‘custom audiences’ ad feature by connecting disparate domains (Zynga’s need for ‘whales’ and user data). Jack Dorsey’s superpower was design, defined as removing friction so thoroughly that no manual is needed, which extended to Square’s risk model shifting to the transaction level.
Effective CEO Communication
Copied to clipboard!
(00:43:39)
  • Key Takeaway: Effective weekly CEO communication should be structured around three dimensions—product, business, and team—with the majority of time dedicated to the CEO’s ’top of mind’ concerns.
  • Summary: A powerful communication artifact is the weekly CEO email, structured into ‘Top of Mind’ (60-70% focus), ‘Performance Update,’ and ‘Miscellaneous’ items like recognition. Leaders should be candid about top-of-mind issues to solicit input from talented employees, as repetition is key for messages to truly resonate. The best North Star metrics balance customer value and business value, avoiding revenue as the sole metric.
Advertising Business Models
Copied to clipboard!
(00:46:34)
  • Key Takeaway: There are only three durable ways to build an advertising business: owning coveted first-party inventory with high intent/identity data, driving specific outcomes at a cost, or being the exclusive provider for a major demand source.
  • Summary: Google and Facebook succeeded by combining high-intent search data with identity data, while AppLovin dominates by delivering the outcome of mobile app installs across the unwashed web. Building a business as a middleman on top of major platforms like Google or OpenAI is doomed, as the platform owners will inevitably learn capabilities and incorporate them. The primary threat to established ad networks is consumer behavior shifting to agentic interfaces that bypass owned apps, reducing engagement opportunities.
Ad Monetization and Engagement Budget
Copied to clipboard!
(00:53:00)
  • Key Takeaway: Facebook managed ad revenue growth against a fixed engagement budget constraint.
  • Summary: Engagement loss from ads must be quantified by comparing ad-free holdout groups against users seeing ads. Teams must adhere to an engagement budget, meaning revenue goals are checked against the maximum acceptable dip in overall engagement, such as the one enforced between the Newsfeed and Ads teams at Facebook.
Defining North Star Metrics
Copied to clipboard!
(00:54:03)
  • Key Takeaway: North Star Metrics must indicate both customer value and business growth, excluding direct revenue.
  • Summary: A good North Star Metric is directly correlated with customer value and leads the business forward; examples include Square’s GPV and Facebook’s DAUs. These metrics require coupling with check metrics, such as gross margin or customer retention, to act as guardrails against optimizing the North Star metric to the detriment of overall company health.
Self-Serve Product Philosophy
Copied to clipboard!
(00:56:00)
  • Key Takeaway: Self-serve capability forces superior onboarding and opens product aperture to millions.
  • Summary: The mandate at Google was to ensure all tools built for internal teams serving large customers were available to small, self-serve customers, which led to faster adoption by smaller users. Self-serve forces intense focus on onboarding and achieving a quick moment of delight, and it allows products to infiltrate organizations from the bottom up, as seen with Figma displacing Sketch.
Thriving Careers in AI Era
Copied to clipboard!
(00:54:50)
  • Key Takeaway: Future success requires becoming a functional expert managing AI agents, not managing humans.
  • Summary: CEOs should hire ‘doers’ and builders over middle managers, as AI agents will automate much of the work currently done by management. The most relevant skill is becoming a functional expert who knows how to build and orchestrate an army of AI agents to perform that function effectively.
Assessing Candidates Through Work
Copied to clipboard!
(01:03:00)
  • Key Takeaway: Work projects are the best way to assess agency and customer voice in candidates.
  • Summary: Engineering excels at using coding interviews, but all functions need work projects to test actual capability, such as analyzing an acquisition target for CorpDev. The best product manager candidates demonstrated agency by rejecting the premise of a proposed product after talking to customers, proving they prioritize the voice of the customer over simply executing assigned tasks.
Career Longevity and Impact
Copied to clipboard!
(01:05:15)
  • Key Takeaway: Job hoppers who leave within 12-18 months are a major red flag for hiring managers.
  • Summary: It takes a minimum of three to four years to have a meaningful impact at a company, so job optimizers doing short stints are doing a disservice to their long-term prospects. People who stick around to build value are the ones who get hired, as short-term thinking prevents the creation of substantial value.
Assessing Founder Superpowers
Copied to clipboard!
(01:06:22)
  • Key Takeaway: Founder authenticity, rooted in lived experience, reveals their core superpower.
  • Summary: The founding story is crucial as it expresses why the founder chose the problem, ideally touching on their inherent superpower and compulsion to solve it. Founders who start companies merely because they want to start a company with a friend lack this authentic drive, unlike those like Figma’s Dylan who are seeped in the domain they are trying to improve.
Navigating the Idea Maze
Copied to clipboard!
(01:08:39)
  • Key Takeaway: Great founders study industry history to justify their chosen solution path.
  • Summary: Founders must articulate how they navigated the ‘idea maze,’ explaining why their chosen solution is superior to alternate approaches. The best founders are students of history in their industry, understanding why prior companies succeeded or failed in tackling the same problem, similar to how the Collison brothers studied payments history before founding Stripe.
Board Composition and Role
Copied to clipboard!
(01:09:33)
  • Key Takeaway: Effective boards require diverse expertise, including product/tech and customer representation.
  • Summary: CEOs should join a board to better understand executive dynamics, and boards should include product/tech expertise, which was rare 15 years ago but is now essential. A ‘board buddy’ system, where board members mentor management team members between meetings, ensures continuous engagement and leverage of board expertise.
Customer Acquisition and Selling
Copied to clipboard!
(01:12:42)
  • Key Takeaway: Enterprise selling must lead with outcomes, leveraging lighthouse customers within specific verticals.
  • Summary: For consumer companies, scaling influencers (especially on platforms like TikTok) is key, while enterprise sales must shift from product features to outcome-based selling, exemplified by Palantir’s ‘solve it or don’t pay’ model. Sales efforts should target one or two specific verticals to secure a lighthouse customer, which then creates a clear path to winning the rest of that vertical.
Gratitude and Paying It Forward
Copied to clipboard!
(01:15:23)
  • Key Takeaway: Gratitude for good fortune compels one to pay forward opportunities without expectation.
  • Summary: Gokul Rajaram cited being hired by Bob McDonald despite being seemingly unqualified as a pivotal, kind act that led him to adopt a pay-it-forward approach. This mindset stems from recognizing the immense fortune of being healthy and in a privileged position, creating a responsibility to help others without expecting anything in return.