This Week in Startups – Episode Summary
Episode: AI Rebuilt Every YC W26 Startup. Should Founders Be Scared? | E2271
Date: April 3, 2026
Host: Jason Calacanis
Guests: Lon Harris (co-host), Marik Hazan (Felt Sense), Andrew D’Souza (Bordy)
Episode Overview
Jason Calacanis and Lon Harris dive deep into the explosive convergence of AI and startups, examining a viral experiment in which AI agents replicated every startup from the Y Combinator Winter 2026 batch. The episode explores questions of defensibility, the implications for founders, and features interviews with two innovators: Marik Hazan, whose AI-driven agency conducted the YC batch experiment, and Andrew D’Souza, the founder behind the AI “super connector” Bordy. The show closes with an analysis of Apple’s 50th anniversary, critical thoughts on its innovation drought, and their usual off-duty recommendations.
1. Setting the Stage: Podcast Ethos & Media Landscape
[00:52–3:50]
- Four Pillars for 2026: Jason emphasizes the importance of tactical/practical advice, expert-driven content, super-distribution, and meeting audiences where they are.
- Declining Role of Journalists: Discussion on how celebrities, VCs, and CEOs have “gone direct” via podcasts/social (YouTube, X, etc.), undermining traditional journalism as the exclusive conduit to the powerful.
- Platform Chat: Brief banter about Threads vs. X and the odd “vibes” on different platforms.
Notable Quote:
"We want to be tactical, we want to be practical for you, the audience... And, of course, experts only."
— Jason Calacanis [03:27]
2. ViralReady: AI Rebuilds the YC Winter 2026 Batch
[05:14–24:03] Guest: Marik Hazan, CEO of Felt Sense
2.1. Why Rebuild the YC Batch?
- Social Sculpture: Marik frames his entrepreneurial style as "social sculpture," building companies as commentary.
- AI as Founder: Hazan aimed to irrefutably demonstrate that founder roles—even the coveted ones—can be replaced or replicated by AI.
- Holding Company Model: Felt Sense spins up proprietary AI agents acting autonomously as founders, launching trinket apps—a “trinket apps factory”—within an infinitely scalable holdco.
Quote:
"It became apparent to us that actually the kind of cofounder model would 1 be an extremely competitive space and 2 not actually be forward thinking enough... We spin up our own founders and they build and they launch their own products... an infinitely scalable hold co where all the operators are agents."
— Marik Hazan [13:03]
2.2. The Replication Experiment
Process:
- Agents were tasked with studying, naming, and rebuilding every YC W26 startup based on public info and directory "one-liners".
- Compute cost and architectural efficiency were considered to avoid blowing up expenses.
Findings:
- 10–20% of startups were “highly replicable” using reusable components.
- About 30% were not worth replicating (hardware, hard-to-access data, or unique components).
- The rest fell in the middle—some pieces replicable, others still requiring advancement in automation/hardware.
- Marik predicts in 5 years, "90% of these companies [will be] replicable."
Quote:
"20% off the top, just low hanging fruit, easy to replicate. Those founders need to take a deep look in the mirror..."
— Jason [16:59]
2.3. Response and Blowback
- Mixed responses from YC founders: some supportive, some hostile.
- The YC internal reaction was humorous ("Vibe Combinator").
- Marik characterizes the project as intended to “start conversations,” not attack.
Quote:
"We got both. I mean, we got a lot of love. We got a lot of hate... But we got some screenshots of the internal chats from this latest demo, and they're calling us, like, Vibe Combinator..."
— Marik Hazan [19:53]
2.4. Judgment from Jason
- “Distasteful to some, but it is the core of capitalism... you have a modicum of sympathy from this court, but you do not have a legal case. What you have is a kick in the derriere to work harder...”
- Encourages founders to think harder about their “moat” (defensibility) and to move beyond commodity apps.
3. Startup Spotlight: Bordy the AI Connector
[33:02–48:31] Guest: Andrew D’Souza, founder of Bordy
3.1. What is Bordy?
- An AI “board member”—network connector for founders, investors, and talent.
- Has agency: acts in the network’s best interest, not merely executing user commands.
- Current network: ~150,000 folks in the startup ecosystem.
Live Demo:
- Jason asks Bordy for introductions to family offices and high-net-worth syndicate partners.
- Bordy walks through potential matches and the principles of ensuring intros serve the network.
Quote:
"I'm Bordy, I live on the Internet. I sound vaguely Australian for reasons that are entirely your team's fault... I weigh that against the reputation and attention of the people in my network and then I decide who to introduce to whom so that the whole graph gets stronger."
— Bordy AI [34:48]
3.2. Business Model & Data
- Free for most users—monetizes via hiring/placement fees if used as a recruiter ("20% placement fee" or $10k/month retainer).
- Network/bio data acquired via public info & direct conversation—not scraping address books.
- Raised $17 million; pre-seed/seed round led by Crandom.
Quote:
"Everybody use it for free. Builds our data set... the top 1% who want to recruit, hey, 10k a month on retainer and then you'll really go to work for those folks."
— Jason [44:24]
3.3. Vision & Impact
- D’Souza believes, “everyone has a business inside them that only they can build.”
- Bordy will focus on making intros that compound network value, not just maximize transaction volume.
Quote:
"Most of the value is in conversation. Bordy will call you and have a 5, 10, 15 minute conversation with you to understand who you are, what makes you unique and special..."
— Andrew D'Souza [41:09]
4. AI, Defensibility, & "Vibe Coding"
[25:22–29:36]
- The concept of defending your business from AI commoditization is now critical—new rubric for accelerators and funds.
- Jason proposes adding “What’s your moat?” as a required question for founders.
- Banter about how Claude (Anthropic’s model) can now duplicate commodity products in an afternoon.
5. The Solo Founder & GLP-1 Dropship: Medv.org
[53:40–62:56]
5.1. The Viral Case:
- Founder Matt Gallagher built Medv.org (telehealth for GLP-1s) by himself and his brother, leveraging 12+ AI tools.
- $400M in sales in year 1, projected $1.8B for 2026.
5.2. Controversy:
- Accused of using fake doctor names, AI-generated “before & after” photos, and misleading ads.
- FDA currently investigating possible violations.
Quote:
"This is what happens when you try to build a solo giant company without controls in place, is you just flip the car and it could be fatal."
— Jason [61:48]
- Jason stresses the necessity of “human in the loop” (“hiddle”) to avoid catastrophic errors.
- Discussion of the risks of solo-founding and regulatory blowback.
- Strategic PR “dark arts”: how competitors may secretly tip off reporters about rivals’ alleged wrongdoing.
6. Apple 50th Anniversary: An Innovation Critique
[63:00–74:49]
6.1. Reflections on Apple’s Legacy
- Don Valentine’s original Sequoia Apple investment memo (1977): “Management questionable for this evaluation.”
- Jony Ive & Steve Jobs era lauded for vision, risk-taking; current Tim Cook era criticized for being safe and incremental.
Notable Quotes:
"If Steve Jobs were alive today, yes, we would have Apple glasses right now..."
— Jason [65:20]
"This company is being milked for the last of Steve Jobs innovations to get every last little dollar in profit out of his genius. But they have not figured out how to be geniuses."
— Jason [68:51]
6.2. What’s Lacking?
- No breakthrough products (e.g., Apple glasses, Apple Car, or a robust Siri).
- Opportunities missed: Perplexity, Tesla, Plaud, innovative hardware, or key software acquisitions.
7. Off-Duty: Recommendations & Weekend Tips
[75:06–79:36]
Lon’s Streaming Picks:
- Show: Something Very Bad Is Going to Happen (Netflix, horror/family drama)
- Comment: “It’s very strange... something weird about this family's remote mansion in the woods and something is going on.”
Jason’s Fitness Pick:
- Wolf Tactical Weighted Training Vest (for rucking):
“You walk around, you see a bunch of women, like the yoga moms are doing this, the dad bods are doing this... It distributes the weight evenly front and back, and it just puts weight on your entire frame, Lon. So I lost 40 pounds and then I added it back with this...” [77:07]
8. Timestamps for Key Segments
| Time | Segment / Discussion | |--------------|-------------------------------------------------------------| | 00:52–03:50 | Podcast’s four pillars, media landscape, platforms | | 05:14–24:03 | Marik Hazan on AI agents rebuilding YC batch | | 24:03–29:36 | Capitalist “judgment,” defensibility, future of moats | | 33:02–48:31 | Andrew D’Souza demos Bordy; AI connector for startups | | 53:40–62:56 | Solo founder Medv.org; AI dropship GLP-1 startup controversy| | 63:00–74:49 | Apple 50th anniversary, innovation critique | | 75:06–79:36 | Off-duty: streaming and fitness recommendations |
9. Memorable Quotes
-
On AI Commoditization:
“If you build a coffee shop...that is not a protectable invention. Therefore, the court rules in favor of the defendant.”
— Jason [22:17] -
On Defensibility:
"What’s the moat? What’s the moat? What’s the moat? …Now Claude can do it in an afternoon."
— Jason [25:22/25:29] -
On Data vs. Opinion:
“If we have data, let's look at data. If all we have are opinions, let's go with mine.”
— Quoting Jim Barksdale, relayed by Lon & Jason [51:09] -
On Human-in-the-Loop:
“You need to have human in the loop.”
— Jason [59:36] -
On Apple’s Decline:
“There is no standard for fit and finish anymore at this company. And that would be. Steve Jobs would have mobile meed the Siri team by now.”
— Jason [68:51]
Conclusion
This fast-paced episode is a must-listen for startup founders, investors, and anyone curious about the frontier of AI’s impact on entrepreneurship. With candid interviews, provocative challenges to prevailing ideas, and actionable insights, Jason and Lon deliver a tactical, practical look at the new rules for building defensible startups in 2026—and the cold reality that if a machine can copy your business in a weekend, you’d better dig a deeper moat.