Generative Now | AI Builders on Creating the Future

PART 2: Matthew Hartman | The Value of Premium Content and the Shifting Economics of the Internet

Date: November 21, 2024
Host: Michael Mignano (Partner, Lightspeed Venture Partners)
Guest: Matt Hartman (Founder & Partner, Factorial Capital)

Episode Overview

In this sequel episode, host Michael Mignano and guest Matt Hartman delve into seismic changes wrought by AI on the internet’s content economy and business models. They explore AI-native SaaS products, the economics and challenges of personalized content, the potential end of ad-driven models, and the rising value of premium content amid content-overload. Honest, exploratory, and occasionally provocative, the conversation is a must-listen for anyone curious about AI, media monetization, and the future of internet economics.

Key Discussion Points & Insights

1. Vertical SaaS, AI-Native Companies, and New Business Models

• AI-Native SaaS: Matt is exploring which B2B SaaS categories are ripe for value creation by new AI-native entrants vs. existing players. He considers the team structure needed and the "private equity" style approach of acquiring legacy businesses to inject AI for cost structure improvements.
  • "Do you start a startup? Or do you buy a company that is doing it an old way and inject AI to change the cost structure?" (Matt, 00:41)
• Service-as-Software: Offering a seamless service where AI powers delivery behind the scenes rather than traditional SaaS products.

2. The Power and Cost of Personalization

• Newsletter Personalization Experiment: Matt describes an RSS-based, AI-powered newsletter generator (nysignal.com) that lets users shape their news feeds based on very specific personal prompts.
  • "With most content platforms...they're ultimately optimized around ad revenue...that's why we're addicted to TikTok. It's junk food. It's not broccoli. But if you had a product that let you actively tune the personalization...you'd have to fund it a different way." (Michael, 03:17)
• Cost of Scaling Personalization: True, user-level AI personalization is expensive to run, especially at scale; inference costs multiply with more users and more granular prompts.
  • "If we have 100,000 users, there's a real cost...That doesn't scale well." (Matt, 04:39)
• Tech Implications: If inference costs drop to (or near) zero, business model possibilities widen:
  • "Right now, all of our assumptions are...the cost of running inference comes down, but it's non-zero. But what if it is fully zero? What does that change?" (Matt, 02:24)

3. Tokens, Business Model Inversion & User Funding

• People Will Pay for Value: Some AI products (e.g., Suno) charge users because there's an evident cost per generation/use, making user funding more intuitively accepted.
• Tokens as Business Model Shift: The way users pay for tokens or computation shifts expectations and could lead to fundamentally new economics for internet businesses.
  • "Tokens, I wonder, are they inverting a business model in some way?" (Matt, 06:18)

4. The End of Advertising (?!) and Internet Economics

• Rise of Answer Engines & Declining Ad Revenue:
  • "We're moving more and more towards a future where agents are going out and capturing the information we want...This is disruptive to the existing business model of advertising in the Internet, which is what makes the entire Internet free." (Michael, 07:02)
• AI Scraping & Tollbit: Michael announces investment in "Tollbit," a company that blocks AI scrapers and makes them pay for access—a sign that content economics might shift from attention (ads) to access (pay-per-use).
• Can Ads Adapt?:
  • "It's easy to be like, well, ads will get figured out. Like ads will be injected into Perplexity...I don't think it's that simple." (Michael, 08:46)
  • "If all of a sudden I had ads inside my Trello, that would be quite annoying." (Matt, 09:06)

5. Content Glut & the Premium Content Opportunity

• Explosion of Content: As creating content becomes trivial, attention becomes ever-more valuable, shifting user and creator dynamics.
  • "If the cost of coding goes to zero and the cost of creating content goes to zero...there's going to be even more noise...the value of someone's attention is that much higher." (Matt, 09:53)
• Premium Content’s Rising Value: Top-tier, differentiated content ("broccoli" vs. "junk food") is likely to become more valuable and sought after.
  • "I just think the value of premium content goes up when you increase the supply." (Michael, 10:49)

6. Subscriptions, Freemium Models, and Future Incentives

• Subscription/Freemium as Survival: As answer engines (like Google's AI Overview) usurp clicks and traffic, the incentive to create free content sharply declines for most internet creators.
  • "For most people on the Internet...the two main incentives for why you would make anything, distribution and revenue, are effectively gone." (Michael, 13:46)
• Creators’ Dilemma & the Long Tail: OpenAI and similar companies may only need the best data, not infinite depth—further driving up the value for top voices, while lessens incentives for long-tail creators.
  • "There's this moment in time...all the Internet up till 2025 looked like X, and then people stopped publishing blog posts because they weren't useful or monetized." (Matt, 17:02)

7. The AI-Powered Creator Economy & Extreme Niche Software

• Zero-Cost Software and the Proliferation of Niche Products:
  • Matt shares building "My Request Room," a hyper-niche, no-code app for piano bar song requests, built in 8 hours, now used by 100+ piano players.
  • "No startup would ever invest in this...but if the cost of building is effectively free...you could have the best version of software for any use case on the Internet." (Michael, 20:06)
• Empowering Non-Tech Creators: With AI and no-code tools, anyone close to a problem or community can create tailored solutions and even charge for them, bypassing the need for venture backing or mass-market scale.

8. The New Golden Age for Product Managers?

• AI Accelerates PM Value: Non-technical founders & PMs can now turn ideas into products directly—no need for deep engineering skills.
  • "Does AI actually make it go back in the other direction? Where if you're a PM or non-technical founder, it's just like—boom—I could just make stuff."* (Michael, 24:06)
• Interface & Interaction Innovation: Building user interfaces for AI remains a mostly unsolved challenge and fertile ground for product innovation —"ChatGPT is MS-DOS, but what is Windows?" (Matt, 24:59)

9. Prompt Engineering, Platform Risk & Defensibility

• Prompt Engineering’s Role:
  • "I was surprised to learn...an AI product that is very beloved...90% of the magic is prompt engineering." (Michael, 26:11)
• Rappers & Platform Dependency: The term "wrapper" is often used derogatorily to describe products with weak defensibility built on top of foundational AI models/API, but this is just modern platform risk—an age-old phenomenon.
  • "To discount an entire category of thing because they use an open API seems quite naive to me." (Matt, 28:53)

Memorable Quotes

• On AI Disrupting Ad Models:
  • "We're moving more and more towards a future where agents are going out and capturing the information we want...This is disruptive to the existing business model of advertising in the Internet, which is what makes the entire Internet free."
    — Michael Mignano (07:02)
• On Content Explosion:
  • "We think there's a lot of stuff on the Internet now, a lot of noise now...if it a hundred folds now, all of a sudden the value of someone's attention is that much higher."
    — Matt Hartman (09:53)
• On Value of Premium Content:
  • "I just think the value of premium content goes up when you increase the supply."
    — Michael Mignano (10:49)
• On Business Models & Platform Dependency:
  • "To discount an entire category of thing because they use an open API seems quite naive to me."
    — Matt Hartman (28:53)

Notable Moments & Timestamps

• [03:17] Michael on content platform addiction and why ad-funded personalization drives "junk food" experiences.
• [06:14] Discussion shifts to user willingness to pay for purposefully burned cloud compute (tokens).
• [07:02] Michael outlines the "end of advertising" argument—answer engines and the future of the web.
• [09:53] Matt highlights the "attention is all you need" idea—content glut strengthening the value of time and attention.
• [13:46] Michael predicts disappearance of distribution/revenue incentives for the "average" creator.
• [20:31] Matt’s case study: building "My Request Room" for piano bar performers as a paradigm for zero-cost, highly personalized software.
• [24:06] Michael and Matt reflect on a new era empowering non-technical creators and PMs.
• [28:53] Matt reframes the criticism of "wrappers" as simply a matter of platform risk, a longstanding issue in tech innovation.

Conclusion

Michael and Matt’s dynamic exchange illuminates both the promise and complexity of the AI-powered future for content, creators, and monetization. From emerging business models to existential threats to advertising and the new creative class enabled by zero-cost software, this episode offers a prescient look at how the engine of the web is being re-wired—and which builders will thrive as the rules shift.

Episode recommended for: AI founders, product managers, content creators, media execs, investors, and anyone following the evolving economics of the Internet.