your system language is:English

Gary Tan on Y Combinator, San Francisco & the Future of Tech

Cover

📺 Today’s recommended deep-dive video: https://www.youtube.com/watch?v=rEwK7MIQ-QA


The YC Way: Gary Tan on Scaling Innovation and Fighting for the Future of Tech

Gary Tan, the President of Y Combinator, shares his journey from a Winnipeg-born engineer to the leader of the world’s most influential startup accelerator. He explores the critical intersection of founder psychology, local politics, and the AI revolution that is currently reshaping Silicon Valley.

Core Question: How can founders balance personal growth and high-speed execution while navigating an increasingly complex regulatory and technological landscape?

Highlights

  • The “Three Eras” of Y Combinator and why it’s returning to “Founder Mode.”
  • Why California’s proposed “billionaire tax” is actually a dangerous asset seizure that could kill the “golden goose” of tech.
  • The psychological trap of “self-abandonment” and how founders can use therapy to improve their leadership.
  • Gary’s “metaprompting” workflow and why the next generation of CEOs must be hackers.

⏱️ Reading time: approx. 8 minutes · Saves you about 76 minutes vs. watching.

Want to take notes while watching? Click the image below and let AI Notebook capture the key points for you 👇

AI Notebook


The Evolution of the Startup Factory

From Summer Internship to Global Powerhouse

Y Combinator didn’t start as a multi-billion dollar institution; it began as a humble experiment by Paul Graham and Jessica Livingston to see if providing small stipends—originally just $13,000—could help hackers build something valuable.

The first batch in Cambridge, Massachusetts, was legendary, producing Reddit and Looped, and setting the stage for what Gary describes as the “First Era” of YC. This period was defined by an internship-style intimacy where the founders were directly mentored by the original four partners. As the program moved to California to escape the brutal winters and stay closer to the rising tech scene, it entered a “Second Era” of scaling, marked by the emergence of massive winners like Dropbox and Airbnb. These successes proved the model worked not just once, but repeatedly, leading to the creation of the continuity fund and the expansion of the partner bench.

Today, YC is in its “Third Era” under Gary’s leadership, focusing on a return to “Founder Mode” and radical ownership.

Instead of moving toward a decentralized or remote-first model, Gary is doubling down on the “managed marketplace” of talent by bringing founders back to San Francisco’s Dogpatch neighborhood. The goal is simple but ambitious: move from participating in 20% of billion-dollar companies to 50%. This requires a relentless focus on the core mantra—”Make something people want”—while providing a high-pressure environment where founders learn to move at an elite speed.

A Gantt chart and timeline showing the three eras of YC evolution: Era 1 (2005-2010) focus on small stipends and direct PG mentorship; Era 2 (2011-2021) focus on scaling, continuity funds, and the rise of Dropbox/Airbnb; Era 3 (2022-Present) focus on Gary Tan's "Founder Mode," in-person batches in San Francisco, and the AI-heavy 30-50% target reach.

💡 Digging Deeper

Q: Why did YC move back to San Francisco and mandate in-person attendance?
A: The remote experience wasn’t producing the same network effects. Being physically surrounded by 800 other capable founders creates a unique energy and a “default trust” that accelerates product-market fit.

Q: What is the current financial deal for joining YC?
A: YC provides $500,000 in exchange for roughly 7% of the company, though the real value lies in the access to the partner guild and the alumni network.

Q: How selective is the program today?
A: It remains incredibly competitive, with a roughly 1% acceptance rate out of approximately 80,000 applications per year.


The Battle for San Francisco

Navigating Policy and the “Asset Seizure” Tax

Silicon Valley is currently facing a political crisis that Gary believes could permanently drive the next generation of founders out of California.

The most pressing threat is a proposed ballot proposition often called the “billionaire tax,” which Gary rebrands as an “asset seizure tax.” Unlike traditional income tax, this proposal targets net worth, but it uses a specific legal “fine print” that calculates ownership based on voting control rather than equity. For founders like Larry Page or Sergey Brin, who hold super-voting shares, this could result in a tax bill that exceeds their entire liquid net worth, effectively forcing them to surrender control of their companies to the state.

A trillion dollars in personal net worth has already left California as billionaires and founders flee to more favorable jurisdictions.

This exodus doesn’t just hurt the wealthy; it guts the tax base that funds middle-class services. Gary argues that tech leaders have been too passive for too long, allowing “anti-tech” supervisors to propose absurd regulations, such as a ban on R&D labs in San Francisco’s Mission District. He views his role now as both an investor and a community organizer, fighting to ensure that the city that birthed the internet, the iPhone, and OpenAI remains a viable home for innovation.


The Internal World of the Founder

Breaking the Cycle of Self-Abandonment

Gary is remarkably open about his own “quarter-life crisis” and the psychological hurdles he had to overcome to lead effectively.

Early in his career, Gary struggled with a pattern he calls “self-abandonment,” where he would suppress his own valid opinions to avoid conflict with co-founders. He mistakenly believed that a “good relationship” meant never fighting, whereas he now understands that great co-founders fight all the time—they just do it while preserving the relationship. This bottled-up frustration eventually led to “rage-quitting” roles at Palantir and even his own startup, Posterous, because he hadn’t learned to communicate his needs in real-time.

Through coaching and therapy, Gary adopted the “Horse and Rider” mental model to gain control over his reactions.

The “rider” is the logical prefrontal cortex, while the “horse” is the emotional, somatic animal that reacts to stress. When founders ignore their “horse,” it eventually rebels, leading to burnout or organizational blow-ups. Gary now teaches founders that their startup will inevitably take on their own personality flaws; if a founder cannot handle conflict, the entire organization will become dysfunctional and conflict-avoidant.

A concept map titled "The Founder's Psychology." At the center is 'Founder Mode.' Branching off are: 'Healthy Conflict' (preserving relationship while debating), 'Self-Abandonment' (the trap of silence leading to rage-quitting), and the 'Horse and Rider Model' (Prefrontal cortex logic vs. Somatic emotional response).

💡 Digging Deeper

Q: How does Gary handle high-pressure moments now?
A: He uses a technique from therapy called “freezing motor functions” to introspect. If his face gets hot, he stops to identify the “if statement” in his emotional code before reacting.

Q: What makes someone a “resilient” founder in Gary’s eyes?
A: The ability to survive a “mortal ego wound.” About a third of YC founders were rejected at least once before getting in; those who keep building despite rejection are the ones who succeed.

Q: What is the most common reason YC startups fail?
A: They don’t usually run out of money; they get “demoralized” or die due to unaddressed co-founder conflict.


The AI Revolution and “Metaprompting”

Why Every CEO Must Be a Hacker

We are entering an era where software will touch a much larger percentage of global GDP, and Gary believes “hacker CEOs” are best positioned to capture this value.

Gary’s personal workflow revolves around “metaprompting”—using AI to write the prompts that then run his business processes. He iterates on these prompts like code; his YouTube scriptwriting prompt is currently on version 27. By feeding his AI tools (Claude, GPT, and Perplexity) examples of his past writing and asking them to introspect on the “pros and cons” of different outputs, he can scale his creative output without losing his personal voice.

The barrier to entry for creating “sanded down,” beautiful products has vanished thanks to tools like Claude Code.

Gary predicts that the next generation of successful companies will be led by people who don’t just “manage” AI, but who are deep in the trenches using it to build. He even suggests that a “long-short fund” based on whether or not a CEO is a practitioner of AI tools would likely outperform the S&P 500. For Gary, the “Red Pill” of the modern era is taking the leap to become a founder and using these tools to see the “real world” of market demand.

A process map of Gary Tan's "Metaprompting" Workflow. Step 1: Input raw notes/ideas into Claude. Step 2: Request an Agentic Prompt based on inputs. Step 3: Compare outputs from 4 LLMs. Step 4: Ask Claude to synthesize the best parts. Step 5: Evolve the base prompt to V+1.


Key Takeaways

The success of a startup is as much about the founder’s internal software as it is about their external product. Gary Tan emphasizes that the ability to process emotions, handle conflict, and avoid “self-abandonment” is a competitive advantage. Founders who invest in their own psychological resilience are less likely to experience the “demoralization” that kills most companies before they hit their stride.

On the macro level, the “Golden Goose” of Silicon Valley is under threat from short-sighted political policies. The battle over San Francisco’s taxation and regulatory environment is a warning for tech hubs everywhere. Founders must become politically engaged to protect the ecosystems that allow high-risk, high-reward innovation to flourish.

Finally, the AI revolution is shifting the definition of leadership. The most successful CEOs of the next decade will be “hackers” who use LLMs to automate knowledge work and iterate at speeds previously thought impossible. By mastering “metaprompting” and staying close to the “real” of the market, founders can build products that truly make something people want.


Q&A

Q1: What does “Make Something People Want” actually mean in practice?
A: It means starting with the verb—making—and then ensuring that the product is built in direct response to human interaction and demand, rather than in a vacuum.

Q2: What was Gary’s first job at YC?
A: He actually started as a “volunteer photographer” and designer, helping founders with their homepages and branding before being promoted to an investing partner.

Q3: How does YC help with fundraising specifically?
A: YC condenses the fundraising process into a two-week “sprint” around Demo Day. They teach founders to treat a “safe” as a starting gun, not a finish line, focusing on growth and momentum to attract investors.

Q4: What is “Bookface”?
A: It is YC’s internal social network and software platform, which Gary himself coded in Rails. It acts as a private directory and knowledge base for the YC “guild.”

Q5: Why does Gary mention the “Matrix” movie as relevant to founders?
A: He views becoming a founder as “taking the red pill.” It allows you to escape the cubicle/corporate “matrix” and engage directly with the “real world” of users and market forces.

Q6: What is the significance of “Founder Mode” for a 20-year-old organization like YC?
A: After a company (or incubator) hits product-market fit, it often falls into “manager mode,” which is about absence and delegation. Gary is re-implementing “Founder Mode,” which is about presence, detail, and radical ownership of the human experience.

Q7: How does Gary use Perplexity vs. Claude?
A: He uses Perplexity for anything requiring web search and synthesis of multiple sources, while he finds Claude to be the most “astute” at evaluating and refining creative writing and prompts.

Leave a Reply

Your email address will not be published. Required fields are marked *

Related Posts