Annotations (18)
“Your ambition has to be to replace the entire system. Every vertical has either a legacy or somewhat new system of record. In 2024, things changed. These companies started seeing that AI agent companies are treating them like a dumb database. So you started seeing last year that these companies are cutting off access to APIs. Slack has done it most publicly. They cut off access to Glean.”— Gokul Rajaram
How to be Durable in the AI Era · p. 5
Strategy & Decision Making · Business & Entrepreneurship
DUR_ENDURING
System of record holders weaponize APIs against agents
“When we were going to launch AdSense in 2003, Sergey was a sponsor. He said, what are you guys building here? We're like, website publishers are going to apply from all across the world. We have to review them and approve them to run AdSense. He's like, why do you need to approve them? We were like, what do you mean? Our ads are going to be running on these things, ads powered by Google. You don't want to be on a porn site. He's like, why not? We didn't even have a good answer. He was right.”— Gokul Rajaram
Great Product Design is Reducing Friction · p. 12
Strategy & Decision Making · Operations & Execution
DUR_ENDURING
Lazy approval: check at scale threshold not upfront
“The software companies that should be the most worried right now is where they are pricing the product based on utility. Zendesk is a good example. Zendesk prices seats, and each seat comes with utility. Each seat corresponds to a customer service agent that takes a certain number of customer tickets. That company should be worried because I can have an AI agent sit right next to Zendesk and siphon off.”— Gokul Rajaram
The Legacy Business Model Ripe for Disruption · p. 6
Business & Entrepreneurship · Strategy & Decision Making · Economics & Markets
DUR_ENDURING
Seat-based pricing vulnerable to partial automation
“Square's whole premise was removing friction from small businesses applying for financial services. When you apply to a bank for payment processing, most small businesses are denied by banks. Square instead said, we are going to accept 90 plus percent of people. But what they did was they put risk at the transaction level. So they accepted you as a person, as a business.”— Gokul Rajaram
Great Product Design is Reducing Friction · p. 11
Strategy & Decision Making · Business & Entrepreneurship
DUR_ENDURING
Shift risk control from application to transaction level
“You want to target a high-value workflow. You want to target a workflow that is deep, that is complex, and that requires custom data. One of the challenges with this whole space is that the models are becoming so good that if you try to build a company that is light, that is not a hard problem, the foundation model companies are going to eat you. I met with the CIO of a Fortune 500 company who said, I don't know why I would use any of these startups.”— Gokul Rajaram
How to be Durable in the AI Era · p. 4
Strategy & Decision Making · Business & Entrepreneurship
DUR_ENDURING
Simple workflows get commoditized by horizontal tools
“Products were deterministic. 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, or what is called evals in AI. Someone has to evaluate whether or not what the software is producing is reasonable or not across various use cases. Who owns the evals? It's the PMs.”— Gokul Rajaram
The Changing Face of Product Development · p. 2
Technology & Engineering · Operations & Execution
DUR_ENDURING
Non-determinism shifts PM role to evaluation owner
“The first thing we are seeing now happen is that PMs are starting to check in code with either Codex or Claude Code, into the actual production repository. Right now, engineers have to review the code, but you're going to soon see that Claude Code, Codex, and other tools actually review the code itself before engineers commit. All the companies are struggling with how to evaluate these people. Earlier, there was nothing called a prototyping interview.”— Gokul Rajaram
The Changing Face of Product Development · p. 2
Operations & Execution · Leadership & Management
DUR_ENDURING
Interview structure follows capability change: add prototyping
“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
The Legacy Business Model Ripe for Disruption · p. 6
Strategy & Decision Making · Business & Entrepreneurship
DUR_ENDURING
Timeless data protects better than ephemeral data
“Jack called the product manager role product editor. Why? Because he believed rightly so that the role of the product manager is not to add more features. Any of us can look at a product and say, here's 10 things you should build. The best designers, the best product people edit down things. We have a hundred features. What are the two things that really matter that will drive the customer outcome? The best designers really will take 10 pages of design and say, cut out all the extraneous stuff.”— Gokul Rajaram
Great Product Design is Reducing Friction · p. 12
Creativity & Innovation · Leadership & Management
DUR_ENDURING
Product editor, not product manager: subtract not add
“There are three fundamental ways to succeed in the ads business. Three, and only three. One, you need to own a very coveted group of users and you need to have a surface on which those users interact. Google Search is a great example. Facebook, we knew who these users were and we could match them to customer data. So you could precisely target these people. ChatGPT, their combination of intent and identity data is unparalleled. Google had intent data but not identity.”— Gokul Rajaram
Three Ways to Build an Ad Business · p. 14
Business & Entrepreneurship · Strategy & Decision Making
DUR_ENDURING
Three ads models: own inventory, drive outcomes, exclusive demand
“The product manager and designer role are merging increasingly. When given the choice between an extra designer and extra engineer, companies are saying, the design systems are already laid out. Now that we have the design systems already laid out, we can use AI to do work around these design systems. So we need maybe a small number of designers at the company level to manage the design systems and the design language, but AI can leverage the design language to do designs.”— Gokul Rajaram
The Changing Face of Product Development · p. 2
Leadership & Management · Operations & Execution
DUR_ENDURING
Designer:engineer ratio collapsing from 1:10 to 1:20
“Outcomes are what define the best product people, and outcomes have to be defined in the form of customer behavior. Because customer behaviors are leading indicators for every business outcome. 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? I tell every CEO, you need to ask why. The only question you need to ask is why. Why are you launching this feature?”— Gokul Rajaram
The Changing Face of Product Development · p. 3
Business & Entrepreneurship · Leadership & Management
DUR_ENDURING
Outcomes = behavior change. Ask why until hypothesis clear.
“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 a thousand AI engineers writing code? You have the big challenge of AI slop. Every product director 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
How to be Durable in the AI Era · p. 4
Leadership & Management · Strategy & Decision Making
DUR_ENDURING
Judgment becomes scarce when production becomes abundant
“Product development as we know it is changing. If you try to be very strict and stringent about describing exactly what you're going to build or prescribing what you're going to build, it will not work. Almost everybody has gone to a bottoms-up approach. Product managers only articulate what the customer needs are at the highest level, and they are the guardian of the why.”— Gokul Rajaram
The Changing Face of Product Development · p. 2
Leadership & Management · Operations & Execution · Technology & Engineering
DUR_ENDURING
PM becomes guardian of why, not what or how
“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 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
How to be Durable in the AI Era · p. 5
Strategy & Decision Making · Business & Entrepreneurship
DUR_ENDURING
Five durability sources: scarce asset, control point, hardware, essential workflow, network effects
“Stay at every job long enough to have impact. Over the last 18, 24 months, I've been seeing this phenomenon of job hoppers who stay at a job for 12 to 18 months and then move to the next job. That is one of the biggest red flags as a hiring manager. I don't think you can achieve anything of value. You can't have any impact on a company in 12 to 18 months. I think it takes minimum three to four years to have impact at a company.”— Gokul Rajaram
Building an Impactful Career · p. 20
Leadership & Management
DUR_ENDURING
Impact requires 3-4 years minimum
“A product person's job is to balance customer needs and business needs. There has to be somebody at the company who's the keeper of the why. Why are we building it? What customer need are we solving? Why is this a pain point? How intense is it and how deep is it? And second, how does it add value to the company? You can build something amazing that adds tremendous value to the customer but doesn't build any value to the business.”— Gokul Rajaram
The Changing Face of Product Development · p. 3
Business & Entrepreneurship · Strategy & Decision Making
DUR_ENDURING
PM balances customer value and business value
“Over the last two months, December 2025 and January 2026, something has fundamentally changed. The notion of a long horizon, long-running agent. About six months ago, I tried to use Claude Code to build a video transcription tool. It kept failing. I had to debug it. I gave up. Two weeks ago, in one hour, I was able to prompt my way to a good video transcription tool. These agents now are resilient to failure, and you don't have to be very technical to use them.”— Gokul Rajaram
The Changing Face of Product Development · p. 1
Technology & Engineering · Operations & Execution
DUR_CONTEXTUAL
6-month capability jump: failure to one-hour success
Frameworks (4)
Bottoms-Up Product Development in High-Velocity Environments
Product organization structure for rapid capability change
When underlying capabilities (tools, models, platforms) change faster than planning cycles, traditional top-down product specification fails. This framework restructures product development around (1) PMs as guardians of customer needs (the why), (2) cross-functional teams working directly on code together, (3) evaluation as the key PM responsibility, and (4) bottoms-up building where the how emerges from capability exploration rather than upfront specification.
Components
- PM owns the WHY, not the WHAT or HOW
- Cross-functional teams work on code together
- Evaluation becomes core PM responsibility
- Hands-on prototyping required
Prerequisites
- Willingness to blur role boundaries
- Access to rapid prototyping tools
- Culture that tolerates failed experiments
Success Indicators
- PMs checking in code to production repositories
- Faster iteration cycles on new capabilities
- Product decisions informed by what's possible, not what was planned
Failure Modes
- PMs becoming junior engineers instead of maintaining strategic view
- Loss of customer focus in favor of technical capability exploration
- Lack of evaluation rigor leading to quality degradation
Behavior-Driven Product Development
Using customer behavior change as product hypothesis
Every product feature must articulate a hypothesis of customer behavior change. The hypothesis answers: if we build this, customers will change from doing X to doing Y. This framework forces product teams to think in terms of outcomes (behavior) rather than outputs (features), creating a testable prediction that can be measured post-launch.
Components
- Define the current behavior state
- Define the target behavior state
- Articulate the mechanism
Prerequisites
- Ability to measure customer behavior
- Data infrastructure for tracking key behaviors
- Culture that accepts hypothesis failure
Success Indicators
- Every feature launch has associated behavior hypothesis
- Post-launch reviews discuss behavior change, not feature adoption
- Failed hypotheses lead to feature changes or deprecation
Failure Modes
- Measuring vanity metrics instead of real behavior
- Changing the hypothesis post-launch to match results
- Not having baseline measurement of current behavior
Five Sources of AI-Era Durability
Building defensibility when software is easy to create
When AI makes software creation nearly free, durability comes from five sources: (1) ownership of scarce assets (licenses, regulations, unique data), (2) control points (money flows, data flows), (3) hardware that's hard to replace, (4) essential workflow positioning, and (5) network effects. A durable business must have at least one, preferably multiple sources.
Components
- Scarce Asset Ownership
- Control Point Capture
- Hardware Integration
- Essential Workflow Positioning
- Network Effects
Prerequisites
- Honest assessment of which sources are actually available
- Willingness to make tradeoffs (e.g., hardware increases CAC)
- Capital to invest in building durability before it pays off
Success Indicators
- Churn rate declining as switching costs increase
- Pricing power improving over time
- Competitors struggling to replicate your position
Failure Modes
- Claiming durability sources you don't actually have
- Building on top of another company's control point
- Choosing the wrong durability source for your market
Three Paths to Advertising Success
The only viable business models in ad-supported businesses
All successful advertising businesses fall into exactly three categories: (1) Own first-party inventory with coveted users (Google Search, Facebook, ChatGPT), (2) Drive guaranteed outcomes at fixed cost (AppLovin for mobile installs), or (3) Be the exclusive provider for large demand sources (The Trade Desk for major brands). Any ad business outside these three paths will fail or be marginalized.
Components
- Path 1: Own First-Party Inventory
- Path 2: Drive Guaranteed Outcomes
- Path 3: Exclusive Demand Aggregation
Prerequisites
- Honest assessment of which path is achievable for your company
- Capital requirements vary dramatically by path (Path 1 requires most)
- Different talent profiles required for each path
Success Indicators
- Clear differentiation from competitors on same path
- Margin expansion as scale increases
- Advertisers unable to replicate results elsewhere
Failure Modes
- Trying to straddle multiple paths
- Building on top of Path 1 platforms without Path 2 or Path 3 strategy
- Lack of exclusivity in Path 3 approach
Mental Models (8)
Platform Commoditization Risk
Strategic ThinkingWhen building on horizontal platforms, anything simple enough to be built by the platform's customer
In Practice: Gokul described CIO saying he'll just use Gemini/ChatGPT to build agents internally, killing startup
Demonstrated by Leg-gr-001
Moat Taxonomy
Strategic ThinkingCompetitive advantages fall into distinct categories, each with different durability and defensibili
In Practice: Gokul enumerated five specific sources of durability, each a different type of moat
Demonstrated by Leg-gr-001
Standardization Enables Leverage
EconomicsOnce a system or process is standardized, it can be leveraged by lower-skilled labor or automation. Design systems standardize visual language, allowing AI to generate designs within the system. The ratio of designers to engineers collapses from 1:10 to 1:20 because standardization happened first.
In Practice: Gokul explained how design systems enabled AI design work, reducing need for designers
Demonstrated by Leg-gr-001
Balancing Dual Optimization
EconomicsMany roles require optimizing two conflicting variables simultaneously. Product managers balance customer value and business value. Optimizing only one destroys the other: pure customer focus bankrupts the company; pure business focus (raising prices) destroys customer base. The skill is finding the frontier where both improve together.
In Practice: Gokul defined product management as balancing customer needs and business needs
Demonstrated by Leg-gr-001
Interview-Job Matching
Decision MakingThe principle that interview processes must directly test for capabilities required in the role. When job requirements change (PMs now must prototype), interview structure must change (add prototyping interview). Failure to match creates systematic hiring errors.
In Practice: Gokul described how companies are adding explicit prototyping interviews because PM role now requires hands-on coding ability
Demonstrated by Leg-gr-001
Abundance Shifts Scarcity
EconomicsWhen one resource becomes abundant, scarcity shifts to complementary resources. When AI makes code production nearly infinite, scarcity shifts to judgment about what code to write. When content creation is free, curation becomes valuable. The pattern: abundance in X creates scarcity in decide-what-X.
In Practice: Gokul identified judgment as the truly future-proof skill because AI makes production abundant
Demonstrated by Leg-gr-001
Data Half-Life
TimeData decays at different rates depending on type. Transactional data (ERP, CRM)
In Practice: Gokul contrasted Slack (short half-life) with NetSuite/Salesforce (long half-lif
Demonstrated by Leg-gr-001
Impact Accumulation Time
TimeMeaningful impact in any system requires minimum time threshold. Cannot have org
In Practice: Gokul warned against job hopping, stating minimum 3-4 years needed for impact
Demonstrated by Leg-gr-001
Connective Tissue (2)
Rick Rubin's Producer as Reducer
Rick Rubin describes his role not as producer but as reducer. His value comes from subtracting elements from recordings, not adding them. This connects directly to Jack Dorsey's reframing of the product manager role as product editor, where the core responsibility is cutting features and design elements, not adding them. The parallel reveals a universal creative principle: in mature creative domains, the scarce skill is editorial judgment (what to remove), not generative ability (what to add).
Gokul mentioned Rick Rubin after describing Jack Dorsey's product editor concept, explicitly connecting the two as examples of the same principle
Hamilton Helmer's 7 Powers
Hamilton Helmer's 7 Powers framework (Scale Economies, Network Effects, Counter-Positioning, Switching Costs, Branding, Cornered Resource, Process Power) provides the foundation for evaluating business durability. Gokul's five sources of AI-era durability map onto Helmer's powers: scarce assets = Cornered Resource, control points = Switching Costs, hardware = Switching Costs, essential workflow = Process Power, network effects = Network Effects. The connection shows how Helmer's framework remains relevant but requires adaptation for contexts where software creation becomes nearly free.
Gokul explicitly referenced Hamilton Helmer's 7 Powers when discussing durability, noting that companies need a few of those powers embedded in their business model
Key Figures (4)
Jack Dorsey
4 mentionsCo-founder and CEO of Square and Twitter
Gokul worked with Jack Dorsey at Square.
- Jack called the product manager role product editor
Sergey Brin
2 mentionsCo-founder of Google
Rick Rubin
1 mentionsMusic producer
Jim McKelvey
1 mentionsCo-founder of Square
Glossary (1)
evals
DOMAIN_JARGONEvaluations; systematic testing of AI system outputs for quality and correctness
“You have to be on the other side, an evaluation, or what is called evals in AI.”
Key People (1)
Hamilton Helmer
Author of 7 Powers: The Foundations of Business Strategy
Concepts (4)
long-running agent
CL_TECHNICALAI system that persists across multiple interactions, maintaining context and handling complex multi-step tasks
control point
CL_STRATEGYStrategic position in a system where money or data flows must pass
network effects
CL_ECONOMICSEconomic phenomenon where product value increases as more users join, creating defensibility
system of record
CL_TECHNICALAuthoritative data source for a business domain; single source of truth for critical data
Synthesis
Synthesis
Migrated from Scholia