Mar 17, 2025 · 33m · tbpn
The GREAT VC AI DEBATE between Slow Ventures and Greylock Partners.
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Venture capitalists Seth Rosenberg and Sam Lessin engage in a high-stakes debate over whether generative AI will overwhelmingly enrich Big Tech incumbents or launch a new generation of venture-scale startups. Grounded in their shared Facebook background and a playful Highway 101 billboard rivalry, they dissect platform defensibility, labor automation, and enterprise software dynamics.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The hosts as informed peer 3.6 Guest teaching 3.3 Guest disagreement 5.2 The hosts pushing back 2.9
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 16:27 Sam dismantles accounting rollups and labor automation theses
Sam aggressively rejects Seth's premise on professional service automation, calling VC accounting rollups stupid and forcefully emphasizing that human jobs exist for liability shifting.
Hardest push from the hosts ▶ 12:10 Host challenges Sam's thesis using TikTok and viral AI contentHost John Coogan directly challenges Sam's insistence that no new networks can be built on commodity AI by citing TikTok's initial lip-syncing wedge and Balenciaga viral AI memes.
Biggest teaching moment ▶ 18:44 Sam dismantles the VC myth of mobile wave disruptionSam educates the hosts and Seth on why the VC narrative of recurring tech-cycle reshuffling is marketing fiction, arguing that mobile and cloud primarily entrenched existing tech giants rather than breaking them.
The host holds their own ▶ 18:21 Host applies Clayton Christensen framework to mobile incumbentsHost John Coogan demonstrates deep structural knowledge by invoking Clayton Christensen's sustaining versus disruptive innovation framework to challenge whether mobile truly disrupted legacy SaaS.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| Pregame Strategy and Highway 101 Billboard Rivalry | 3 | 1 | 2 | 1 | The hosts and Seth engage in friendly pregame banter regarding the physical billboard war on Highway 101 between Slow Ventures and Greylock. Host John Coogan sets up the debate framework and jokes about Sam filibustering. | |
| Seth Rosenberg Reflects on Facebook Origins with Sam | 4 | 2 | 1 | 2 | Seth recounts his early career at Facebook after being hired by Sam, reflecting on early messenger bot hype cycles. The host demonstrates domain knowledge by interjecting with observations on Project Titan, bot logic, and hyperscaler PM dynamics. | |
| Sam Lessin Joins Stream and Frames Debate Premises | 3 | 2 | 5 | 2 | Sam Lessin joins the stream and aggressively frames his central thesis that AI will enrich big platforms while startups lose money. Seth counters by identifying content democratization and marketplace opportunities. | |
| Content Democratization and the Pipeline Distribution Debate | 3 | 3 | 7 | 4 | The co-host pushes Seth on whether creator platforms like Suno will inevitably leak distribution to existing platforms like YouTube. Sam passionately rejects the idea of new media networks, using an information fracking metaphor to argue all value flows to existing pipelines. | |
| VC Probability Math and AI Integration in Merit First | 5 | 4 | 6 | 5 | Host John Coogan challenges Sam by comparing AI video generation to the early lip-syncing hook that birthed TikTok. Sam dismisses venture capitalists for doing sloppy probability math averaging zero and infinity before detailing Merit First. | |
| Labor Automation Versus Liability Shifting in Professional Services | 2 | 6 | 8 | 1 | Seth argues AI will replace human labor spend in compliance and professional services, but Sam forcefully intervenes. Sam schools the room on why clients hire human accountants for liability shifting to the IRS rather than computational accuracy. | |
| Mobile Wave Parallels and Incumbent Resilience Dynamics | 6 | 5 | 7 | 6 | The host pushes Sam on Christensen's innovator's dilemma framework and whether mobile created distinct venture outcomes like WhatsApp. Sam rejects the premise, labeling VC cycle-reshuffling claims as marketing mythology while Seth defends mobile-first outliers. | |
| Sclerotic Industries and the Vanishing Middle Market | 5 | 4 | 6 | 4 | The host questions if IT consultancies like Accenture can be disrupted by AI startups or if VC as an asset class is broken. Sam explains that the middle market evaporates, warning that subscale venture-backed startups are bringing knives to gunfights against hyperscalers. | |
| Developer Tooling, Zero Switching Costs, and Synthetic Networks | 3 | 4 | 7 | 3 | Seth offers Cursor as proof of AI startups achieving rapid distribution, but Sam dismisses developer tools as lacking defensibility due to zero switching costs. The panel discusses how AI bots are inverting social network mechanics. | |
| Enterprise SaaS Dynamics and the Collapse of Outbound Workflows | 4 | 5 | 7 | 3 | The co-host raises Tome's pivot and Salesforce's aggressive bundling tactics. Sam argues that AI-generated outbound emails will entirely destroy the value of email inboxes, rendering that entire software category obsolete. | |
| Concluding the Debate and Scheduling the Five-Year Verdict | 2 | 0 | 1 | 1 | The hosts ring the gong to end the debate, joking about bot-driven audience voting and scheduling a five-year calendar reminder to check the outcome of both investors' bets. |
Statements from this episode (21)
Rosenberg: Lessin rented a billboard declaring AI is not your moat
“He actually took out a physical billboard on the one-on-one like with all, like amongst all of these like advertisements for AI companies that says, AI is not your moat. Like love slow kind of thing.”
Rosenberg: Messenger bots failed because LLMs were seven years away
“So I was running the product for the developer platform at Facebook, which was like AI bots, messenger bots. It's like the hottest thing in Silicon Valley for like four months before everyone realized it didn't work because we were seven years away from LLMs a…”
Lessin: AI will enrich Big Tech platforms but burn startup investors
“AI is going to make the big, big, big platforms a shit ton of money. And it's going to be really good for small businesses, but it is a terrible place to hunt for startups, right? It's a great way to lose a lot of money on startups.”
Rosenberg: Massive AI-native businesses will emerge from content democratization marketplaces
“I think massive new businesses can be built that are, let's say, AI native. I think one is like, networks and marketplaces based on new democratization of being able to create new types of content.”
Rosenberg: Suno could build a closed ecosystem if content quality scales
“But I think if like the number and quality of things that are actually created on the Suno platform starts to exceed, or at least like take a meaningful, but separate piece of the market, then they could just create a more closed ecosystem. Where the only way …”
Lessin: Flawed expected-value math causes VCs to overpay for AI startups
“The problem with AI in general, and I think the entire AI discussion is I always like to point out, it's like the average of infinity and zero is infinity, right? And so VCs aren't very good at math, but they can do that one. And so the problem is you say, oka…”
Lessin: Winning investments back transformed workflows, not the underlying AI
“It's not, you're not investing in the AI. You're investing in like a new way or a new approach to hiring, which happens to leverage AI. And I think there is an important distinction there.”
Lessin: Many human jobs exist primarily to shift liability, not execute tasks
“You don't understand, like, this is a liability shifting thing. Like most, a lot of human jobs are not actually about doing the thing. They're about shifting liability and responsibility, right?”
Lessin: Selling AI tools to professionals is a commoditized war of attrition
“And like, that's not where the leverage is. So you're not really investing in the AI business, right? You're still investing in the accounting business, Steve's business, right? And everyone has the same level that it's a classic war of attrition where like yo…”
Lessin: AI is a sustaining innovation that primarily enriches incumbents
“Whereas I could not think of a more classically extending innovation than AI, right? It just makes the rich richer.”
Lessin: Startups exploiting incumbent execution errors will be copied, not acquired
“Like they're not going to get acquired. They're just going to get copied.”
Lessin: Accenture will cannibalize its core BPO business with AI
“Accenture is going to make a lot of money short-term on trans on this stuff, right? Short-term. The problem is Accenture is mostly in the business of BPO outsourcing and they're going to kind of eat themselves in doing that.”
Lessin: Being a $10B public company is a dumb place to be
“There's no such thing as being a ten billion dollar public company anymore. It's just like a dumb place to be. No one cares, right? You either want to be a hyperscaler, right?”
Lessin: AI leverage will create one-person billion-dollar companies
“I totally buy the thesis that you're going to see individuals that have a lot of trust and brand and community and deeply understand a niche. Do things you haven't seen before. Like, I buy the one person billion dollar company thing. Like, I think that's real …”
Lessin: Cursor is a great product but not a good business
“Cursor's, I like it. I use it as a product. I don't see it as a good business.”
Lessin: AI drives software switching costs to zero, destroying value
“I think the whole point of AI is the switching cost for almost all this stuff goes to zero, right? And because switching cost is zero, there's just no value.”
Lessin: New AI social networks focus on choosing who follows you
“For 20 years, the way you start a social network is you say, who do you want to follow? And the way to start a social network today is say, who do you want to follow you?”
Lessin suspects Twitter intentionally allows bots to satisfy user loneliness
“I'd actually argue Twitter is already doing this. Like Twitter, I think is intention. My, I'm not, I have no inside information. My sense is it is intentionally full of bots. Cause people are lonely. People want responses.”
Lessin: Tome cannot compete and outbound sales industry will collapse
“No way can Tom compete on this, but I'll go even a step further, which is this entire industry of outbound goes away.”
Rosenberg: Tome will succeed without needing to defeat Salesforce
“Look, I don't think you need to be beat Salesforce to be very successful and to find a segment of the market that wants a more modern product that, you know, understands your business has everything in one place and can actually make your sales team a lot more…”