Hollywood is about to hit a "James Cameron" moment where $200M visions no longer require $200M budgets.

In this episode, I sit down with Mehul, the 24-year-old founder of Koyal (YC), who is building the platform that generated the first full-length Bollywood film entirely with AI. We dive into why traditional experience is becoming a liability in Silicon Valley, the "Film Arena" benchmarking the world's best models, and why the future of Netflix is a "choose your own adventure" generated in real-time.

What you’ll learn:

- Why Silicon Valley views 24 as the cutoff for "young" founders.

- The "Film Arena": How Koyal benchmarks Runway, Luma, and Kling.

- Why the "Infinite Canvas" UI is a mistake for creative tools.

- How AI allows for "reshooting" a movie without a single set.

- The future of interactive media: Starring in your own Harry Potter.

📌 Watch now and subscribe for more smart, fast, founder-first interviews.

TIMESTAMPS:

00:00 AI Films Explained

00:33 Sponsor YC Roaster

01:13 Meet Koyal Founder

02:01 Young Founder Advantage

02:58 Experience Versus New Tech

07:01 Co Founding With Family

10:46 Democratizing Movie Budgets

12:35 Actors Avatars Hybrid Sets

14:02 Selling To Hollywood

15:46 Koyal Workflow Like Filmmaking

18:30 Hollywood Adoption And Theaters

20:37 T Series And Product Evolution

25:07 Shooting Impossible Locations

26:35 AI Beats Green Screen

26:46 Avatar Lesson on Timing

27:52 AI Augments Creators

28:56 Future Media Two Paths

29:32 Real Time Interactive Movies

30:50 Bandersnatch and Engagement

34:30 Pixels to 3D Immersion

38:09 How Much Control Is Too Much

39:19 Personalized Harry Potter Worlds

41:29 Licensing IP and UGC

42:42 Koyal Strong Product Opinions

48:37 Wrap Up and Community Plans

🎧 Listen on the go:

Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP

Apple: https://apple.co/4cZ8RMm

My other YouTube Channel: https://www.youtube.com/@lobstercap

This Lobster Talks episode is brought to you by YC Roaster, a free tool helping founders improve their YC applications.

🔗 https://ycroaster.com

Most developers are still coding in the past.

While the world is obsessed with AI autocomplete, a new breed of YC founders is building a future where one developer manages ten AI agents in parallel.

In this episode, I sit down with Kiet, co-founder of Superset (YC P26), to discuss the massive shift from writing code to orchestrating swarms of agents. We dive into why traditional IDEs like Cursor are failing the agentic era and how "worktrees" are the secret to scaling your output.

What you will learn:

• Why the role of a developer has shifted from "Writer" to "Editor" to "Orchestrator."
• The technical limitations of current IDEs when running multiple agents.
• How to use Git worktrees to give every AI agent its own isolated environment.
• The future of "Intent" and how tools like Linear will integrate directly into the code.
• Why "human-in-the-loop" is the current meta, and where it goes next.

📌 Watch now and subscribe for more smart, fast, founder-first interviews.

TIMESTAMPS:

00:00 Cursor Becomes Secondary

00:51 YC Roaster Ad Break

01:35 Podcast Welcome And Setup

02:41 Meet Superset And The Problem

05:12 Why Cursor No Longer Fits

07:13 Superset Parallel Workspaces

10:32 Go To Market And Growth

14:50 Staying Relevant In Fast AI

16:29 Invariants And Agent Future

20:53 Building Primitives Now

24:13 Linear And Task Systems Debate

27:12 APIs And Defensibility

28:41 YC Shoutouts And Whats Next

30:11 YC Records And AI Boom

31:43 AI Adoption Gap Outside SF

32:30 YC Cycle Compresses With Agents

10:32 Should YC Be One Month

35:50 Second Time In YC

37:18 Why Repeat Founders Return

41:13 YC Confidence And Mentorship

46:00 Leaving A Startup To Start Superset

48:08 Team First Thesis For Superset

53:11 Risk Taking And Fear Setting

56:49 Superset Plug And Farewell

🎧 Listen on the go:

Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP

Apple: https://apple.co/4cZ8RMm

My other YouTube Channel: https://www.youtube.com/@lobstercap

This Lobster Talks episode is brought to you by YC Roaster, a free tool helping founders improve their YC applications.

🔗 https://ycroaster.com

There is a dangerous "founder aura" in Silicon Valley… the idea that you must be a 22-year-old Ivy League dropout with a perfect linear trajectory to win. It’s a myth that keeps great founders from starting.

Conor and Manu didn’t check a single "consensus" box: they were in their 30s, didn't come from FAANG, and were rejected by YC multiple times. Yet, Connor became the "star" of his batch by doing the opposite of what most PR firms advise: he leaned into the mess. He posted about sleeping in closets, being neurodivergent, and the "vibe" of his team.

Authenticity isn't just a "nice to have" for your personal brand; it’s a filter for your cap table. By being brutally honest about the grittiness of the journey, you repel the "fair-weather" investors and attract the ones who actually understand that startups are mostly a series of controlled fires.

In this episode, we sit down with Conor from Hyper Spell (YC F25) to discuss the "star" founder's journey, the brutal reality of YC rejections, and why the future of sales belongs to agents, not humans.

What you’ll learn:

• Why Hyper Spell refunded all its customers to pivot just months before YC.
• How to optimize your technical documentation for "Agentic Discovery."
• The "Slope vs. Intercept" framework for hiring wizards over Ivy League grads.
• Why authenticity and "vibe" are the ultimate trust-builders in Silicon Valley.
• The "Agent-to-Agent" hiring process that ignores resumes entirely.

📌 Watch now and subscribe for more smart, fast, founder-first interviews.

00:00 Resilience Mindset

00:37 YC Roaster Sponsor

01:21 Podcast Warm Welcome

01:57 Why YC Community

04:58 Hyperspell Explained

05:33 Origin Story Pivot

09:29 Pivot Timing YC

11:03 YC Rejections Lessons

00:00 No Means Not Yet

14:19 Authenticity Builds Brand

15:52 Breaking Founder Myths

19:23 Being Real Online

21:16 Unique Founder Stories

26:22 Twitter and Customers

27:53 Authenticity Builds Trust

29:44 Overdelivering for Customers

32:45 Humans Still Decide

36:44 Selling to Agents Playbook

39:01 Building for Agent UX

43:29 Agent Hiring Meritocracy

48:00 Choosing the Right Cofounder

53:57 Hyper Spell Roadmap

56:01 Closing Thanks and Wrap

🎧 Listen on the go:

Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP

Apple: https://apple.co/4cZ8RMm

My other YouTube Channel: https://www.youtube.com/@lobstercap

This Lobster Talks episode is brought to you by YC Roaster, a free tool helping founders improve their YC applications.

🔗 https://ycroaster.com

Bain Capital just bet $5.6 billion on legacy EdTech. Noah and the Scout team are betting on its extinction.

Most founders are told to avoid the Student Information System (SIS) market because it's "too sticky" and dominated by giants like PowerSchool.

In this episode, Noah (Founder of Scout, YC W25) explains how they are using AI-native compliance to wedge into the "School OS" market and why the future of education looks nothing like the traditional classroom.

What you’ll learn:

- Why Private Equity buying your competitor is a massive "Buy" signal for your startup.

- The "Wedge" Strategy: Why Scout chose online hybrid schools in California.

- How to overcome intense Imposter Syndrome during the first week of Y Combinator.

- The Sneaker Bot Origin: Why YC loves founders who "hacked" systems as kids.

- Managing the "Post-YC Slump" and building a 20-year sustainable business.

📌 Watch now and subscribe for more smart, fast, founder-first interviews.

Chapters:

00:00 YC Investor Boost

00:49 Free YC Roaster Plug

01:34 EdTech Market Shakeup

02:16 What Scout Builds

03:05 PowerSchool Deal Timing

03:48 From Wedge to SIS

06:56 Beating Switching Costs

09:52 Niche Beachhead Strategy

11:00 First Real Customer Signal

13:41 Future of Schooling Bet

16:31 Why Based in New York

19:50 NYC Talent Network Edge

22:11 Sneaker Bot Origins

24:33 Building the Supreme Bot

25:38 Hacking Systems Mindset

26:30 Spotting System Hacks

27:24 Waffles and Tooth Bonds

28:59 Getting Into YC

30:31 Imposter Syndrome Reset

34:23 YC Lessons That Stick

34:50 Going Big in EdTech

39:37 The Post YC Slump

42:54 From Demo Day to Customers

45:42 YC Stamp for Investors

47:21 Rapid Fire and Wrap

🎧 Listen on the go:

Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP

Apple: https://apple.co/4cZ8RMm

My other YouTube Channel: https://www.youtube.com/@lobstercap

This Lobster Talks episode is brought to you by YC Roaster, a free tool helping founders improve their YC applications.

🔗 https://ycroaster.com

The YC Winter 2026 batch just broke every record in the book.

We’re seeing valuations hit $100M before founders even finish their slide decks. Is this a peak or just the new baseline?

In this episode of The Lobster Talks, we debrief the W26 Demo Day. We go behind the scenes on the "Agentic" boom, why hardware is suddenly the hottest play in the valley, and the real math behind why YC startups are worth 4x more than their peers.

What you’ll learn:

• Why 10+ companies in this batch raised at $100M+ valuations.
• The truth about the "1% acceptance rate" (and how to skip the line).
• How a consumer hardware startup hit $27M ARR during the batch.
• Why "Open Claw" changed the game for AI agent startups.
• The blurring lines between Seed and Series A rounds.

📌 Watch now and subscribe for more smart, fast, founder-first interviews.

Chapters:

00:00 Faster Exits Thesis

00:30 YC Roaster Plug

01:14 Winter 2026 Debrief

01:59 Record Growth Metrics

03:17 Revenue And Acceptance

06:44 Why YC Still Pays

07:42 100M Valuation Surge

11:23 Seed Series A Blur

14:49 Skipping Series A

19:47 AI Agents Narrative

22:35 Why YC Still Wins

23:22 Post Claude Demo Day Shift

24:43 Seed FOMO And Valuation Surge

25:40 Why Exits Keep Getting Bigger

27:33 Big Tech Acquisition Logic

34:10 Talent Scarcity Drives M&A

36:37 Breaking Down 27M ARR

38:50 Consumer AI Notetaker Bet

43:00 Wrap Up And Where To Follow

🎧 Listen on the go:

Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP

Apple: https://apple.co/4cZ8RMm

My other YouTube Channel: https://www.youtube.com/@lobstercap

This Lobster Talks episode is brought to you by YC Roaster, a free tool helping founders improve their YC applications.

🔗 https://ycroaster.com

He did YC in 2016 from France with a Slack bot nobody had heard of.

9 years later, he came back… second company, second batch, $10M seed closed in 3 days. Dalton told him: "Don't overlearn."

Quang is the founder of Vybe, a YC W25 company building the internal app layer for the AI era. Before that: Plato, a mentorship platform for engineers he grew to 70 people, raised $22M for, and sold to Coda, where he worked as a GM before quitting to start over.

In this conversation, we go deep on what it actually means to be a second-time YC founder: the overcorrections, the traps, and what's genuinely different about YC today vs. 2016.

In this episode, you'll learn:

→ Why hiring zero people after raising $10M was the wrong overcorrection, and when Quang finally figured that out

→ How selling before the product was ready backfired, and what First Round's Liz told him to do instead

→ YC's self-fulfilling prophecy: how the program became a distribution machine, and why seed rounds that took 3 weeks in 2016 now close in 3 days

→ Why Vybe is quietly pivoting from "build internal apps" to "agents that use those apps", and what inspired it

→ The three-component PMF framework First Round runs with every portfolio company (and where most founders skip straight to step 3)

Chapters:

00:00 Raising First 150K

00:30 YC Roaster Sponsor

01:14 Meet Quang Twice YC

02:03 Discovering YC In France

05:37 Slack Bot Breakthrough

06:34 Product Hunt To YC Interview

07:49 Silicon Valley Network Effect

13:47 Why Expense Reports Question

15:59 Getting The Acceptance Call

17:02 Moving To Mountain View

18:46 Building Bots In YC

20:31 Pivot To Plato Mentorship

22:02 Acquired By Coda

22:49 Coda Vs Retool Pain

26:37 Replit Agent Sparks Vibe

27:33 Internal Apps Are Still Hard

28:52 Quitting Coda For Vibe

29:00 Co Founder Wealthfront Pain

30:07 Back To Y Combinator

30:26 Avoiding Overcorrection Traps

31:48 Hiring Lessons And Rebound

34:33 Selling Too Early Too Hard

37:39 Five Customers Before Scale

39:17 Product Market Fit Framework

42:08 Claude Code Threat Or Tailwind

44:52 Open Claude And Apps Disappear

46:46 Vibe Agents And Integrations

49:10 YC Then Vs Now

55:49 Wrap Up And Subscribe

—

🦞 Lobster Talks is the YC insider podcast for investors. New episodes weekly.

Subscribe and turn on notifications so you don't miss the next one.

Listen on the go:

🎧 Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP

🎧 Apple: https://apple.co/4cZ8RMm

More from Lobster Capital:

📺 YouTube: https://www.youtube.com/@lobstercap

🌐 Website: https://lobstercap.com

This Lobster Talks episode is brought to you by YC Roaster, a free tool helping founders improve their YC applications.

🔗 https://ycroaster.com

The recovery market is broken, and most VCs are too scared to touch it.

Koby Conrad (CEO of Sunflower) joins the show to discuss why he’s betting his career on solving addiction using AI.

After a successful exit at Rupa Health, Koby moved to Buenos Aires with nothing but a backpack to "vibe code" the future of sobriety.

In this episode, we dive into the "no-meeting" SAFE phenomenon, why SaaS moats are dying, and how Sunflower reached $1M ARR in record time.

What you will learn:

• Why the "Wellness" category is a trap for serious founders.
• How to raise capital in "emotionally difficult" or stigmatized markets.
• The strategy behind giving away software for free to monetize physical "atoms."
• Inside the YC Fall 25 batch: What has changed since 2019.
• Why AI sponsors are actually more effective for lonely users than human ones.

📌 Watch now and subscribe for more smart, fast, founder-first interviews.

Chapters:

00:00 Choosing Mission Over Money

00:30 YC Roaster Sponsor Break

01:14 Why Addiction Is Underserved

02:08 What Sunflower Builds

03:24 Fundraising In A Stigmatized Market

05:14 Why Not A Wellness App

08:00 Stigma Parallels And Founder Backstory

10:37 YC Twice And The Compounding Effect

13:59 No Meeting SAFEs And Investor Alpha

18:31 Limits Of AI In Recovery

22:06 From Subscriptions To Clinics

28:14 Distribution Moats And Community

30:27 Community Network Moats

31:21 Software Is Not The Moat

31:57 B2B Pilots And Revenue Paths

33:18 Finding Product Model Fit

34:56 Why Start Over For Addiction

37:01 Backpack Minimalism Reset

39:24 Garry Tan As Group Partner

43:04 Rapid Fire AI Predictions

43:45 Gambling Addiction Rabbit Hole

46:34 AI Across Healthcare And Ash

50:59 Smarter Than The Robot

51:53 Where To Find Sunflower

🎧 Listen on the go:

Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP

Apple: https://apple.co/4cZ8RMm

My other YouTube Channel: https://www.youtube.com/@lobstercap

This Lobster Talks episode is brought to you by YC Roaster, a free tool helping founders improve their YC applications.

🔗 https://ycroaster.com

The US government spends $2.7 trillion a year on everything from water bottles to fighter jets, yet the entire process is held together by Microsoft Word and tribal knowledge.

In this episode, we sit down with August Chen, founder of Hazel (YC W24), to discuss why he left Palantir to tackle the antiquated world of government procurement. We dive into the "missionary vs. mercenary" mindset, the reality of the 1% acceptance rate at YC, and why the "death of the paper-pusher" is the greatest opportunity in VC today.

What you’ll learn:

• Why 50% of small businesses have abandoned government contracts.
• The Palantir playbook: How to "deploy yourself" into a customer's workflow.
• Why selling to the Government requires a "hierarchy of needs" approach.
• The pivot from B2B2G to direct B2G: When the mission stays, but the product moves.
• How AI is turning months of RFP paperwork into seconds of work.
• Predictions for the "DOGE" era and the future of civil servant tech stacks.

📌 Watch now and subscribe for more smart, fast, founder-first interviews.

Chapters:

00:00 Million Dollar Toilet Problem

00:39 YC Roaster Ad Break

01:22 Why Gov Procurement Lags

02:50 Broken Systems Examples

04:05 Inside Procurement Tech Stack

05:30 Palantir Lessons On Procurement

07:29 YC Pivot To B2G

11:36 Selling To Government Reality

17:22 Scaling A Small Sales Team

21:29 AI For Deterministic Compliance

25:15 Workflow First AI Second

26:00 Vendor Search Enrichment

26:18 Drafting Requirements Fast

26:36 Evaluation Notes Assistant

27:04 Safety and Tesla Analogy

28:35 Scope Review Guardrails

30:29 Palantir Patience Lessons

32:36 Hair on Fire Problems

33:57 Fast Pilots and Features

36:07 Overcoming Trust Deficit

39:04 Why the Name Hazel

41:50 Predictions Rapid Fire

43:58 Procurement Jobs Future

45:32 Investor Takeaway and Wra

“🎧 Listen on the go:

Spotify: https://open.spotify.com/show/1u1JyRKH8JYFjhBkvTUOAP

Apple: https://apple.co/4cZ8RMm

My other YouTube Channel: https://www.youtube.com/@lobstercap

This Lobster Talks episode is brought to you by YC Roaster, a free tool helping founders improve their YC applications.

🔗 https://ycroaster.com