Annotations (18)
“Square's whole premise was removing friction from small businesses applying for financial services. Square at its core is a risk company. When you apply it to a bank for payment processing, most small businesses are denied by banks when they apply. Square instead said, we are going to accept 90%+ of people. What they did was they put risk at the transaction level.”— Gokul Rajaram
Strategy & Decision Making · Business & Entrepreneurship · Technology & Engineering
DUR_ENDURING
Shifted risk from approval to transaction level
“Every vertical has either a legacy or somewhat new, what is called a system of record, which is a system where most of the data is stored for that system. For legal, there's a company called Filevine or another company called Clio. In sales, it's Salesforce. In healthcare, it's Epic. For many years, these companies all had APIs that if you enter that industry, you could build an agent company on top of these APIs. In 2025, things changed.”— Gokul Rajaram
Strategy & Decision Making · Business & Entrepreneurship · Technology & Engineering
DUR_ENDURING
Systems of record closing APIs to protect moats
“There are two kinds of legacy companies. One are systems of records, and one are things that are priced based on outcomes. The software companies that should be the most worried right now is where they're pricing the product based on utility. Zendesk is a good example. Literally Zendesk prices seats and each seat comes with utility. Each seat corresponds to a customer service agent that attacks certain number of customer tickets.”— Gokul Rajaram
Strategy & Decision Making · Business & Entrepreneurship · Economics & Markets
DUR_ENDURING
Seat-based pricing most vulnerable to AI disruption
“When I was growing up, products were deterministic where there was a workflow. You knew if a user did X, Y happened. Today you could do X, Y happens, but if you do slight variation of X, something completely different happens. Non-deterministic software. What that means is you have to be on the other side, an evaluation of what is called evals in AI. And someone has to evaluate whether or not what the software is producing is reasonable or not across various use cases. Who owns the evals?”— Gokul Rajaram
Technology & Engineering · Leadership & Management · Operations & Execution
DUR_ENDURING
Non-deterministic software requires AI to evaluate AI
“Custom audiences is a foundation of most ad systems now, is the idea that as an advertiser, you want to reach people who are similar to your customers. If you're a bank and you have 100,000 customers, how can you give this set of customers to your ad platform and say, instead of describing these customers, what did ad systems do before? They would describe their customers. I think they are 25 to 34-year-old women. That's not good enough.”— Gokul Rajaram
Technology & Engineering · Business & Entrepreneurship · Economics & Markets
DUR_ENDURING
Zuck invented lookalike audiences from Zynga request
“Gmail was launched on April 1st, 2003. In the internal alpha it said this gives you 1 gigabyte of storage. Back then, Yahoo Mail was the dominant product and it gave 10 megabytes of storage. So this thing had 100x more storage. This really epitomizes Larry and Sergey's philosophy, which was basically build the best technology on the planet. They were deeply technical and every product was held to technology and scale. AdSense was the fastest growing product in Google history.”— Gokul Rajaram
Strategy & Decision Making · Leadership & Management · Technology & Engineering
DUR_ENDURING
Larry cared about market share, not revenue
“Almost everybody has gone to a bottoms-up approach where it's not done by product management anymore. Product managers, the only thing they do now is they articulate what the customer needs are at the highest level, and then they're the guardian of the why. But the actual product is built bottoms-up by engineers, researchers, and product managers and designers all working together on the code itself. Capabilities and models are changing very fast.”— Gokul Rajaram
Leadership & Management · Technology & Engineering · Operations & Execution
DUR_ENDURING
Product managers become guardians of why, not what
“I think it's going to be very hard for an end customer to use multiple companies where you have a system of record and then you have this agent that sometimes doesn't work with it properly. So the agent companies have no option but to also start building and offering a system of record. So every company I know is now trying to figure out how do I build the entire platform and not just a system that does some workflows. Last year everyone was like, oh, we can do workflows.”— Gokul Rajaram
Strategy & Decision Making · Business & Entrepreneurship · Technology & Engineering
DUR_ENDURING
Agent companies forced to build full platform
“There's explicit interview in the interview loop called prototyping. Literally forces product managers to be hands-on. Second, the product manager and designer role are merging increasingly. A lot of companies are going through headcount allocation this year, and I'm hearing from many teams that when given the choice between an extra designer and extra engineer, they're saying, you know what, the design systems are already laid out.”— Gokul Rajaram
Leadership & Management · Operations & Execution · Technology & Engineering
DUR_CONTEXTUAL
PM to engineer ratio moves from 1:10 to 1:20
“Everything you do or build should be attuned to the goal of what customer state change does it lead to? What customer behavior change does it lead to? You need to ask why. The only question you need to ask is why. Why are you launching this feature? And you should not let any feature go out if there's not a clear hypothesis behind this feature. And the hypothesis has to be articulated in the form of a customer behavior change.”— Gokul Rajaram
Business & Entrepreneurship · Strategy & Decision Making · Leadership & Management
DUR_ENDURING
Every feature must articulate behavior change hypothesis
“I met with the CIO of a Fortune 500 company a few weeks ago. He said, look, I don't know why I would use any of these startups. Gemini has an agent builder product and I also use ChatGPT Enterprise and they also have an agent builder product and I have 1,000 IT engineers who work for me. They all want to be retrained as AI engineers. So I'm just going to put them using these horizontal tools to build my AI agents. Why do you need any startups?”— Gokul Rajaram
Strategy & Decision Making · Business & Entrepreneurship · Technology & Engineering
DUR_CONTEXTUAL
If enterprise can build it internally, no startup moat
“There are a few things around durability. One, you need to have ownership of a scarce asset. A scarce asset could be a license of some kind. It could be a regulation of some kind where you have unique insight into it. Second, you might need to, you might basically own a control point. A control point is a thing that controls how people interact with money or with data. Third, you want to maybe have hardware which is hard to replace. Fourth, maybe you want to be part of an essential workflow.”— Gokul Rajaram
Strategy & Decision Making · Business & Entrepreneurship
DUR_ENDURING
Five sources of durable competitive advantage
“Eric would give a product leader— we would become seconded to Eric for the weekly strategy or the annual strategy planning session. My job was to go to Eric and say, Eric, how do you want to present the strategy of the company? He's like, well, it's very simple. I want you to go and interview each of the different leaders of the different teams. There's only one constraint I have. You can't use any words to describe what they're doing. You have to use only images. I'm like, why is that?”— Gokul Rajaram
Leadership & Management · Psychology & Behavior · Creativity & Innovation
DUR_ENDURING
People remember feelings, not words
“The one thing I think that's going to be truly future-proof is judgment. Why? Because what is the biggest challenge you have when you have 1,000 AI engineers writing code? You have the big challenge of AI slop. Every product leader I've talked to is extremely worried that because you have these engineers running rampant, they're just going to produce lots of code. Which of this code is even valuable?”— Gokul Rajaram
Leadership & Management · Strategy & Decision Making · Technology & Engineering
DUR_ENDURING
Infinite productivity requires finite judgment
“The companies that are less exposed are ones where the utility is not based on seats, but it's based on data that has been collected and captured over a period of time. The more timeless the data is, the more protected they are. ERP is a great example. Somebody uses NetSuite as an ERP. There is no compelling reason for someone to put their career at stake by ripping out NetSuite.”— Gokul Rajaram
Strategy & Decision Making · Business & Entrepreneurship · Psychology & Behavior
DUR_ENDURING
Career risk protects ERP systems from disruption
“In the age of AI, stickiness I think comes from a few sources. One, you need to have network effects. So DoorDash is sticky not just because it has this beautiful app, but it's because it's a network of restaurants and dashers and consumers. The second example of stickiness is when you have financial or money moving through you. Many of the system of records, for example, Toast, have payments going through them.”— Gokul Rajaram
Strategy & Decision Making · Business & Entrepreneurship
DUR_ENDURING
Four sources of stickiness in AI era
“Product managers articulated what to build, designers designed it, and engineers built it. Over the last 2 months, December 25th and Jan 26th, it's become very clear that something has fundamentally changed. About 6 months ago, I tried to use Claude Code to build a video transcription tool. I tried to build it, it kept failing, and then I had to go in and try to debug it. Ultimately I gave up.”— Gokul Rajaram
Technology & Engineering · Business & Entrepreneurship · Operations & Execution
DUR_ENDURING
6-month capability jump marks inflection point
“A product person or product manager, their job is to balance customer needs and business needs. The product manager has to be somebody at the company who's a keeper of the why. Why are we building it? What customer need are we solving? Why is this a pain point? How intense it is, how deep it is. And second, how does this add value to the company? If you build this thing solving this customer need, how does value add to the company? Balancing those two is a very delicate act.”— Gokul Rajaram
Business & Entrepreneurship · Strategy & Decision Making · Leadership & Management
DUR_ENDURING
Product manager as keeper of the why
Frameworks (2)
Customer Behavior Change Hypothesis Framework
Feature justification through behavioral outcomes
Every feature launch must be justified by a clear hypothesis about customer behavior change. The hypothesis articulates what state change the feature will drive, moving customers from doing X to doing Y, or from spending X minutes to Y minutes. This framework prevents feature bloat and ensures product development stays grounded in customer outcomes.
Components
- Identify Current Customer State
- Define Target Customer State
- Articulate the Mechanism
Five Sources of AI-Era Durability
Building defensible moats when software is commoditized
In an era when AI makes software creation trivially easy, durability requires ownership of scarce assets, control points, hardware, essential workflows, or network effects. This framework helps companies identify which sources of defensibility they can realistically build and sustain against both AI-native competitors and incumbent platform companies.
Components
- Scarce Asset Ownership
- Control Point Acquisition
- Hardware Integration
- Essential Workflow Embeddedness
- Network Effects Engineering
Mental Models (12)
Customer-Business Value Balance
Decision MakingEvery product decision sits at the intersection of customer value and business value. Building something amazing for customers that creates no business value is as fatal as extracting business value in ways that destroy customer value. The product manager's core job is maintaining this balance.
In Practice: Discussion of product philosophy and the PM role
Demonstrated by Leg-gr-001
Risk Assessment Level Shifting
Decision MakingWhere you assess risk determines who you can serve. Assess at approval, and you must reject most applicants. Assess at transaction, and you can accept most applicants while maintaining safety. The pattern applies broadly: shift expensive gates to cheap continuous checks to expand addressable market.
In Practice: Discussion of Square's payment processing risk innovation
Demonstrated by Leg-gr-001
Automation-Driven Span of Control Expansion
EconomicsAs AI automates coordination and oversight work, the optimal number of direct reports per manager increases. What was 1:10 becomes 1:20 because the manager's leverage comes from judgment and strategic direction rather than tactical coordination. This changes organizational design fundamentally.
In Practice: Discussion of changing designer and PM to engineer ratios
Demonstrated by Leg-gr-001
Judgment as Scarce Resource in Age of Abundance
EconomicsWhen production becomes essentially free, the constraint shifts to curation and evaluation. As AI makes creating code, designs, and content trivially easy, the bottleneck becomes deciding what to create and evaluating what was created. Judgment becomes the only truly scarce input.
In Practice: Discussion of what skills remain future-proof in AI era
Demonstrated by Leg-gr-001
Seat-Based Pricing Vulnerability
EconomicsSoftware priced by seats (per-user licensing) is most vulnerable to AI disruption because AI agents can directly substitute for human seats. Companies with seat-based models must rapidly shift to outcome-based pricing or risk being hollowed out as customers replace human seats with AI agents.
In Practice: Discussion of why Zendesk is vulnerable to AI disruption
Demonstrated by Leg-gr-001
Career Risk as Switching Cost
PsychologyThe single biggest switching cost in enterprise software is not technical integration complexity or financial cost, it's the career risk to the decision-maker. Ripping out a system that runs the business is career-limiting if it goes wrong, so incumbents with deep data and critical workflows are protected by human loss aversion.
In Practice: Discussion of why NetSuite is insulated from AI disruption
Demonstrated by Leg-gr-001
Emotion Creates Memory, Words Do Not
PsychologyPeople remember how things made them feel, not what was said. Visual and emotional content sticks in memory while purely verbal content evaporates. This is why storytelling beats bullet points and why the best communicators translate ideas into images and feelings.
In Practice: Eric Schmidt's image-only strategy presentations
Demonstrated by Leg-gr-001
Horizontal Platform Cannibalization
Strategic ThinkingWhen horizontal AI platforms become powerful enough, they enable customers to build internally what they previously bought from specialized vendors. This creates an existential threat to vertical SaaS companies that don't have deep moats. The defense is moving from tool to platform or embedding deeper into workflows.
In Practice: CIO discussing why he doesn't need startups when he has Gemini and ChatGPT
Demonstrated by Leg-gr-001
Seven Powers Redux in AI Era
Strategic ThinkingHamilton Helmer's seven powers framework (scale economies, network effects, counter-positioning, switching costs, branding, cornered resource, process power) remains relevant but requires AI-era translation. Network effects and cornered resources become more important as pure software advantages erode.
In Practice: Discussion of sources of durability for AI companies
Demonstrated by Leg-gr-001
API Access as Strategic Weapon
Strategic ThinkingPlatform companies initially offer open APIs to encourage ecosystem development, but once AI agents threaten to commoditize their value add, they weaponize API access through restriction, bundling, or pricing. This forces dependent companies to vertically integrate or die.
In Practice: Discussion of Slack cutting off Glean's API access and broader pattern
Demonstrated by Leg-gr-001
Four Stickiness Sources in AI Era
Strategic ThinkingWhen software becomes trivially easy to create, stickiness comes from network effects, financial flows, hardware integration, or unique asset access. Pure software moats erode rapidly; compound moats require at least two of these four elements working together.
In Practice: Discussion of what creates stickiness in AI era
Demonstrated by Leg-gr-001
Role Boundary Collapse Under Acceleration
Systems ThinkingWhen the rate of capability change exceeds the rate at which organizations can adapt formal structures, role boundaries dissolve and reform around outcomes rather than functions. Product managers who were gatekeepers of what gets built become guardians of why it gets built, while engineers take on traditional PM responsibilities because the tools allow them to.
In Practice: Discussion of how AI is changing product development roles
Demonstrated by Leg-gr-001
Connective Tissue (2)
Square shift from approval-time risk to transaction-time risk assessment
Traditional banks assess risk at the approval stage. Square innovated by accepting 90%+ of applicants upfront and running ML models at each transaction.
Discussion of what Gokul learned from Jack Dorsey about removing friction at Square
Gmail launching with 100x more storage than Yahoo Mail
When Gmail launched on April Fools Day 2003 with 1GB storage, Yahoo Mail only offered 10MB. The 100x improvement was categorical.
Discussion of Larry and Sergey philosophy at Google
Key Figures (12)
Jack Dorsey
4 mentionsSquare Co-founder and CEO
Larry Page
3 mentionsGoogle Co-founder and CEO
Mark Zuckerberg
3 mentionsFacebook Co-founder and CEO
DoorDash
2 mentionsFood Delivery Platform (Company)
Toast
2 mentionsRestaurant POS and Payments (Company)
Sergey Brin
2 mentionsGoogle Co-founder
Eric Schmidt
2 mentionsGoogle CEO
Mark Pincus
2 mentionsZynga CEO
Slack
2 mentionsCollaboration Platform (Company)
Jim McKelvey
2 mentionsSquare Co-founder
Zendesk
2 mentionsCustomer Service Software (Company)
NetSuite
2 mentionsERP Software (Company)
Glossary (3)
evals
DOMAIN_JARGONEvaluations, the process of testing AI system outputs for quality and appropriateness
“What that means is you have to be on the other side, an evaluation of what is called evals in AI.”
AI slop
DOMAIN_JARGONLow-quality output produced rapidly by AI systems without adequate human judgment or oversight
“What is the biggest challenge you have when you have 1,000 AI engineers writing code? You have the big challenge of AI slop.”
whales
DOMAIN_JARGONHigh-spending customers who generate disproportionate revenue, especially in gaming and mobile apps
“80% of all revenue for any gaming company comes from whales.”
Key People (3)
Sergey Brin
(1973–)Co-founder of Google with Larry Page
Larry Page
(1973–)Co-founder of Google with Sergey Brin
Eric Schmidt
(1955–)CEO of Google from 2001-2011, brought operational discipline
Concepts (6)
long-running agent
CL_TECHNICALAI system that persists across sessions, maintains context, and can recover from failures
system of record
CL_TECHNICALAuthoritative data source for a business function
system of action
CL_TECHNICALSoftware layer that performs workflows on top of system of record
ERP (Enterprise Resource Planning)
CL_TECHNICALIntegrated software system that manages core business processes
network effects
CL_ECONOMICSPhenomenon where value of product increases as more people use it
custom audiences / lookalike audiences
CL_TECHNICALAd targeting method where advertisers upload customer lists and platform finds similar users
Synthesis
Dominant Themes
- AI-driven collapse of traditional product development roles
- Judgment as the last scarce resource
- Defensibility requires compound moats
Unexpected Discoveries
- PM to engineer ratio moving from 1:10 to 1:20
- Non-deterministic software requiring AI to evaluate AI outputs
- Slack cutting off Glean as harbinger of platform defense strategies
Cross-Source Questions
- How would Carnegie or Rockefeller think about vertical integration in the API economy?
Processing Notes
Source was rich in transferable principles despite contemporary context.
Synthesis
Source was rich in transferable principles despite contemporary context.