$46B of hard truths from Ben Horowitz: Why founders fail and why you need to run toward fear (a16z co-founder)

Sep 11, 2025 Episode Page ↗
Overview

Ben Horwitz, co-founder of A16Z, discusses the psychological muscle required for leadership, the importance of making unpopular decisions, and how to build confidence as a CEO. He also shares insights on AI opportunities, venture capital philosophy, and the value of supporting pioneering hip-hop artists.

At a Glance
38 Insights
1h 37m Duration
16 Topics
4 Concepts

Deep Dive Analysis

Success as a Series of Small Decisions

The Importance of Running Towards Fear

Who Should (and Shouldn't) Start a Company

Databricks: Thinking Bigger with Investment

Managerial Leverage and CEO Development

When to Replace a Founder as CEO

Normalizing Failure and Building Confidence as CEO

Counterintuitive Lessons in Company Building

The 'Good Product Manager/Bad Product Manager' Essay

Product Managers as Leaders and Mini-CEOs

Investing in Adam Neumann and Judging Strengths

Is AI in a Bubble? A Contrarian View

Biggest Opportunities in AI: Infrastructure and Applications

The Importance of US Leadership in AI

The Paid in Full Foundation for Hip-Hop Pioneers

Lightning Round: Books, Products, and Life Motto

Running Towards Fear

This leadership principle involves making difficult decisions quickly, especially when faced with two undesirable options, rather than hesitating. Hesitation is considered the most destructive action for a leader, as it can paralyze a company and lead to worse outcomes.

Managerial Leverage

This concept describes a situation where direct reports proactively generate ideas and drive their organizations forward, rather than the manager constantly dictating tasks. It signifies that the manager gains more value from the report's contributions than they would if the report wasn't there.

Founder Mode (nuanced)

Initially, the wisdom was to quickly hire senior executives after product-market fit. However, Ben Horowitz clarifies that founders should build their executive team slowly and deliberately, pacing it to their ability to integrate and manage new hires, rather than avoiding experienced hires altogether.

Kin-based Culture

A societal structure where cooperation and knowledge sharing are primarily confined to family units, hindering broader societal advancements like science, cities, and large companies. This contrasts with cultures that historically enforced monogamous marriage, which fostered wider cooperation.

?
What is the key to success in leadership?

Success is a cumulative result of many small, difficult decisions, requiring the psychological ability to let go of sunk costs. Overcoming self-limiting beliefs and not allowing external perceptions to define you is also crucial.

?
Why is it important for leaders to 'run towards fear'?

Hesitation is the most destructive action for a leader, especially when faced with two bad choices. Running towards fear means making the difficult, often unpopular, decision quickly to prevent worse outcomes like bankruptcy or organizational paralysis.

?
Who should not start a company?

Individuals primarily motivated by money should avoid starting companies, as the extreme difficulty and pain involved often outweigh the financial rewards. True founders possess an irrational desire to build something greater than themselves and improve the world.

?
When should a founder be replaced as CEO?

Founders are typically replaced when they lose confidence due to past mistakes, leading to hesitation in decision-making. This hesitation can paralyze the company and foster a political, dysfunctional environment as senior staff attempt to fill the leadership vacuum.

?
What is the core message of the 'Good Product Manager/Bad Product Manager' essay?

The essay emphasizes that a product manager's role is fundamentally one of leadership, requiring the ability to influence and guide teams to ensure the product's success, even without direct authority over them.

?
Why did a16z invest in Adam Neumann after WeWork?

Andreessen Horowitz's investment philosophy focuses on judging individuals by their strengths and potential, rather than solely by past failures. They recognized Adam Neumann's impressive ability to build a globally recognized brand as a significant strength.

?
Is the current AI market in a bubble?

Ben Horowitz suggests it's likely not a bubble in the same way the dot-com era was, as current AI companies demonstrate strong revenue growth and viable business models. However, the technology's early and immature state means current market positions might not be sustainable, potentially leading to future market dislocations.

?
Where are the biggest opportunities in the AI industry for founders?

Significant opportunities lie in infrastructure, particularly in running open-source models with high efficiency and low latency. The application layer also presents vast potential, especially for companies that develop proprietary models or leverage unique data to create deep, defensible moats beyond simple 'thin wrappers' over foundation models.

?
Why is US leadership in AI important?

The US system, despite its imperfections, excels at distributing power and fostering innovation, providing opportunities for all. Historically, nations that led in industrialization gained global power, and AI represents the next wave of industrialization. US leadership is crucial for humanity to avoid dangerous concentrations of power globally.

?
How can leaders build trust in an organization, especially from scratch?

Based on Shaka Senghor's experience, simple actions like ensuring everyone eats together can build rapport. Additionally, establishing and strictly enforcing clear moral codes, such as always being truthful both internally and externally, is vital when starting with individuals who may lack inherent trust.

1. Avoid Hesitation as a Leader

Do not hesitate on difficult decisions, especially when faced with two bad options, as hesitation is the most destructive thing a leader can do, leading to company paralysis and internal political dysfunction.

2. Build Psychological Muscle for Leadership

Develop the ability to confront difficult situations and choose the ‘slightly better’ path, even if both options are horrible, as this psychological muscle is crucial for great leadership.

3. Overcome Self-Limiting Beliefs

Recognize that external negative perceptions or events are less damaging than what you believe about yourself; by rejecting self-defeating beliefs, you can overcome almost any challenge.

4. Run Towards Pain and Darkness

Actively confront scary decisions and difficult conversations rather than avoiding them, as hesitation is destructive and facing the pain directly often leads to better outcomes.

5. Add Value Through Unpopular Decisions

Understand that your unique value as a leader comes from making decisions that most people don’t like or agree with, because if everyone agreed, your leadership wouldn’t be necessary.

6. Break Sunk Costs, Make Small Good Decisions

Psychologically break free from sunk costs to avoid bad paths, and consistently make small, difficult ‘good decisions,’ believing each will lead to the next positive outcome and ultimately success.

7. Maintain Confidence to Prevent Dysfunction

As a founder, maintaining confidence is crucial because losing it leads to hesitation, which can cause senior team members to fill the void, leading to internal politics and a dysfunctional organization.

8. Embrace D-Minuses, Avoid Fails

Be comfortable with making many small mistakes or getting ‘D-minuses’ as a CEO, as long as you avoid catastrophic failures (‘F’s) like running out of cash, and keep moving forward.

9. Prioritize Long-Term Respect Over Short-Term Liking

As a leader, focus on earning long-term respect by telling difficult truths, even if it means people won’t like you in the short term, because this honesty is vital for the company’s survival and success.

10. Hire People Who Make You Great

As a CEO, your primary role in team building is to find and hire world-class talent who can push the company forward and from whom you can learn, rather than trying to develop people in functions you don’t master.

11. Seek Managerial Leverage

Ensure your direct reports are proactively identifying and driving initiatives to advance the company; if you constantly have to tell them what to do, you’re not getting leverage and a change might be necessary.

12. Start a Company for Irrational Desire, Not Money

Only embark on starting a company if you have an irrational desire to build something larger than yourself and improve the world, as financial motivation alone is insufficient to endure the inherent difficulties.

13. Focus on Impact Beyond Profit

Founders should be driven by an abstract desire to create something important, where people enjoy working and benefit from the products, rather than solely by the pursuit of success and wealth.

14. Think Big and Commit Fully

When starting a company, aim for significant impact and commit fully with adequate resources (e.g., raising $10 million instead of $200,000), rather than thinking too small or hedging your bets.

15. Build Executive Teams Slowly and Deliberately

Avoid rushing to hire a full team of senior executives; instead, build the team slowly and at a pace that allows you to effectively integrate and manage them, preventing deference and political infighting.

16. Hire Experienced Talent Strategically

While avoiding blind deference, strategically hire experienced professionals who possess specialized knowledge (e.g., in sales) to accelerate specific functions and prevent costly trial-and-error learning.

17. Lead Through Influence as a PM

Recognize that the Product Manager role is fundamentally a leadership position requiring influence, as you must guide teams to achieve product success without formal authority (no direct reports, firing/promoting power).

18. Be the ‘Mini-CEO’ of Your Product

As a Product Manager, act as the ‘mini-CEO’ for your product, focusing on its ultimate success, consolidating ideas, prioritizing, and ensuring high-fidelity understanding of the vision across the team.

19. Give Specific, Actionable Feedback

When addressing performance issues, provide specific feedback on the gap between current performance and required role, offering support for improvement while clearly outlining consequences if the gap isn’t closed.

20. Improve Explanations to Avoid Yelling

If you find yourself yelling at your team, it likely indicates a failure on your part to clearly explain your expectations or vision, suggesting you need to improve your communication.

21. Judge People by Strengths, Not Past Failures

When evaluating individuals for investment or hiring, focus on their world-class strengths and what they can do, rather than fixating on past mistakes or weaknesses, as everyone is flawed.

22. Coach to Strengths, Not Weaknesses

When coaching or managing, help individuals leverage their strengths and maximize their potential, rather than excessively focusing on or hand-wringing about their weaknesses.

23. Invest in World-Class Strengths

In venture capital, prioritize investing in founders with world-class strengths that can overcome competition, acknowledging that everyone has flaws and planning to surround them with people who can compensate.

24. Leverage Networks for CEO Confidence

Accessing a powerful network from day one can significantly boost a founder’s confidence and ability to navigate challenges by connecting them with key decision-makers and experienced leaders.

25. Seek Support to Build CEO Confidence

Engage with mentors or firms that provide a strong network and peer support (like CEO barbecues) to help you feel capable and maintain confidence as a CEO, especially when facing self-doubt.

26. Distrust Superficial Advice

Be wary of short, generic advice snippets, especially from those without direct experience, as true wisdom in company building is subtle, complex, and requires deep understanding.

27. Focus on Efficient AI Infrastructure

Invest in or build solutions that can run open-source AI models with the lowest cost and latency, as this infrastructure layer will be extremely valuable in the AI industry.

28. Secure Significant Capital for Foundation Models

To compete in the foundational AI model space, be prepared to raise at least $2 billion for training alone, as this is the cost required to achieve competitive capability.

29. Build Complex, Sticky AI Applications

Focus on developing AI applications that are more complex and ‘stickier’ than simple wrappers around foundation models, as these offer significant opportunity and competitive moats.

30. Build Proprietary Models/Data for AI Moats

To create a competitive advantage in AI applications, develop your own specialized models or accumulate proprietary data through user interaction, as this goes beyond generic foundation models.

31. Address Complex Problem Spaces with AI

Recognize that many real-world problems, especially in enterprise settings (e.g., semantic issues, access control, long-tail human behavior), require more than basic foundation models and offer significant opportunity for specialized AI solutions.

32. Prioritize US Leadership in AI for Global Opportunity

Support US leadership in AI development and policy, as a strong US position is crucial for maintaining a system that provides broad opportunity and prevents concentrated power globally.

33. Foster Trust Through Shared Experiences

Build trust and rapport within a team by encouraging shared experiences, such as eating meals together, to reinforce a sense of unity and collective identity.

34. Enforce Strict Honesty and Integrity

Establish and enforce a strict code of honesty and integrity, both internally and externally, as a fundamental principle for building a strong, trustworthy culture, especially when starting from scratch.

35. Accept Life’s Unfairness to Move Forward

Internalize the understanding that life is not fair; letting go of the expectation of fairness allows you to focus on actionable solutions rather than dwelling on injustices.

36. Focus on Action, Not Prediction

As a CEO, prioritize building solutions and taking action (‘building an ark’) rather than merely predicting problems (‘predicting rain’), as only action leads to positive outcomes.

37. Support Pioneering Artists (Paid in Full Foundation)

Consider supporting organizations like the Paid in Full Foundation, which provides financial assistance (pensions) to pioneering hip-hop artists, enabling them to continue their work and receive recognition.

38. Don’t Assume a Bubble if Everyone Says So

A true market bubble requires widespread belief that it isn’t a bubble; if everyone is already calling it a bubble, it’s less likely to be one because capitulation hasn’t occurred.

The only value you ever add is when you make a decision that most people don't like.

Ben Horowitz

The psychological muscle you have to build to be a great leader is to be able to look in the abyss and go, okay, we're going that way slightly better. We're going to go that way.

Ben Horowitz

You don't make people great. You find people that make you great, that make the company great, that you learn from, not the other way around.

Ben Horowitz

You don't judge a person by the worst thing that ever happened to them.

Ben Horowitz

Life isn't fair.

Ben Horowitz's father

There's no credit will be given for predicting rain, only credit for building an arc.

Ben Horowitz's manager

Addressing a CTO's 'Asshole' Behavior

Ben Horowitz
  1. Sit down with the CTO for a direct conversation.
  2. Acknowledge their strengths and contributions (e.g., 'You're a really good director of engineering because you do a great job in managing the team, get the products out, all that.').
  3. Clearly explain the gap in their role as a CTO: 'But, like, you're not really a CTO because to be a CTO, you have to be effective with other parts of the organization. You can't just be, like, effective only with engineering.'
  4. Provide specific examples of the negative impact of their behavior, such as making someone cry, and explain how it undermines their effectiveness with other departments.
  5. Offer support for improvement ('if you want to get good at that, I'll help you all work with you on it') while clearly stating the consequence if no change occurs ('But if you don't, I'm going to have to hire a CTO at some point because, like, obviously I need that.').
$2 million
Loudcloud's revenue at IPO Trailing 12 months revenue when the company went public at 18 months old.
600
Number of people at Andreessen Horowitz Most are focused on building a network for founders.
Almost 500 people
Peak attendance at Ben Horowitz's CEO barbecue Held in his backyard to help founders build confidence.
18
Median score on the 'CEO test' An informal assessment, indicating that CEOs often make many mistakes and need to be comfortable with 'D minuses'.
90%
Netscape's browser market share in 1996 At a time when the internet had only 55 million total users.
55 million
Total internet users in 1996 Half of whom were on dial-up connections.
300 engineers
Engineers at Evite (greeting card company) during dot-com bubble Illustrates the difficulty and cost of building products during that era.
At least $2 billion
Estimated cost to train a competitive foundational AI model Required to achieve competitiveness without significant product progress.
80%
OpenAI's estimated share of AI revenue Indicates the dominance of foundational models in current AI revenue.
Close to 200 nights a year
Rakim's touring frequency before Paid in Full award Highlighting the financial struggles of pioneering hip-hop artists.
Six months
Time between Roxanne Shante's Paid in Full award and Grammy Lifetime Achievement Award Demonstrates the impact of the foundation's recognition.