Big Technology Podcast: "Where Are The AI Startups?"

Host: Alex Kantrowitz
Guest: Rick Heitzmann, Managing Partner, FirstMark Capital
Date: October 15, 2025

Episode Overview

This episode of the Big Technology Podcast delves into the current state of AI startups in an era increasingly dominated by foundational models such as ChatGPT and Claude. Alex Kantrowitz is joined by Rick Heitzmann, a seasoned venture capitalist from FirstMark Capital, to unpack why the wave of consumer-facing generative AI startups has stalled, what sectors still hold promise, and the broader economic and societal implications of mass AI deployment.

Key Discussion Points and Insights

1. The Absence of Consumer AI Startups (00:33 - 07:29)

• Why the Startup Boom Hasn’t Arrived:

  • OpenAI and ChatGPT have created “a really good product with breadth and depth” (02:17) that’s hard for startups to surpass.
  • Many broad consumer use cases, like diet coaching or fitness, are now effectively accessible via general-purpose AI models, reducing the need for specialized startups.
  • Enterprise AI ventures like Harvey (legal), Henry (real estate), and Evolution IQ (insurance) succeed due to access to domain-specific, often private data, enabling them to deliver superior, tailored experiences.

• Data as the Key Differentiator:

  • Consumer AI applications often rely on general web data and lack the differentiated data sets seen in successful enterprise AI startups.
  • “Your AI is only as good as your underlying data and your training data.” (02:30, Rick)

• Point Solutions vs. General AI:

  • Startups attempting to wrap ChatGPT for niche uses (e.g., travel, math tutoring) rarely offer significant improvement to warrant independent investment.

Notable Quote

“We frankly have been a bit frustrated by the lack of startups we've seen in their ability to invest along those lines.”
— Rick Heitzmann, 03:38

2. Is Everything Doomed to Be a "Wrapper" on ChatGPT? (07:29 - 13:21)

• The Challenge of Defensibility:

  • As context windows in LLMs expand, even startups with verticalized offerings (like Harvey in law) may lose their edge.
  • The majority of “rote work at the bottom of the legal pyramid” is vulnerable to broad-based AI platforms.

• The Wrapper Debate:

  • Kantrowitz raises whether all distinct AI applications will simply become “wrappers” for broad platforms, complicating the traditional startup-investment thesis.

Notable Quote

“Does all this stuff end up just happening within the ChatGPT interface?”
— Alex Kantrowitz, 10:23

3. What Remains Investable: The Case for Enterprise Data & Security (13:21 - 15:54)

• Enterprise as the Battleground:

  • Heitzmann sees investable opportunities where startups have access to “defined customers” with “defined sets of data” and can create walled gardens for security and performance.

• Data Privacy/Fear as a Market Force:

  • Rising concerns over data sharing with large language models are driving demand for proprietary, secure, in-house AI applications.

Notable Quotes

“Is that data privacy going to be a key limiter to how the next generation of companies evolve?”
— Rick Heitzmann, 13:18

4. The Role and Limits of "Creative Destruction" (16:03 - 20:24)

• Do We Lose Something Without Consumer AI Startups?

  • Kantrowitz asks if users miss out on the specialization and problem-solving ethos of traditional startups.
  • Heitzmann expresses faith in the “creative destruction” process, stating that humans and experts will always find ways to add unique value on top of AI-driven platforms.

• Personalization and the Human Touch:

  • The next horizon may be highly personalized coaching or assistance augmented by AI and, perhaps, “humans as value-add.”

Notable Quotes

“The human's job is to stay just ahead of that technology and understand where they could create unique and discrete value.”
— Rick Heitzmann, 18:35

5. Real-World AI Applications & Ethical Considerations (20:24 - 22:45)

• Social and Wellbeing Use Cases:
  • AI companions for the elderly, education, and even as tutors are rapidly becoming more lifelike—and valuable.
  • Concerns persist about AI's influence on mental health and the risks of dependency or negative reinforcement.

Memorable Moment

Kantrowitz describes a Korean AI-powered stuffed animal for elderly companionship, which can even check if users are taking their medication (20:45).

Notable Quotes

“AI companionship is an incredible thing … it could be your medical buddy, your math buddy, your surf buddy.”
— Rick Heitzmann, 22:01

6. Is There Too Much Money in AI? The Funding Boom & Risks (23:22 - 31:02)

• The Scale of Investment:
  • Hyperscalers (e.g., Microsoft, Nvidia, Oracle) are self-funding AI infrastructure at unprecedented levels, far exceeding past tech build-outs.
  • The Nvidia–OpenAI relationship exemplifies financial engineering in the sector, with investments described as partnerships rather than traditional VC checks.

Notable Quotes

“Jensen just ... recently committed $100 billion to OpenAI one day, one check.”
— Alex Kantrowitz, 25:51

“The best part … of OpenAI being private is they could do a lot of these deals where they don't have to be disclosed … which also is very much a symbol of a very frothy market.”
— Rick Heitzmann, 27:50

7. Broader Economic Impacts: Will AI Automate All Work? (36:58 - 49:12)

• Automation Across Both White and Blue Collar Sectors:

  • Routine tasks in law, banking, customer service, and even manufacturing are being automated.
  • Heitzmann highlights historical precedents (agriculture, legal, finance) where automation shrank labor demand in old roles but humans created new ones—arguing for long-term optimism.

• Gen Z & Entry-Level Work:

  • Anxiety about the future is high given current employment statistics and the lower number of entry-level positions due to automation.
  • The rise of the creator economy and entrepreneurial pivots offer optimism for the future of work.

Notable Quotes

“Going back to farming … about 93% of Americans were in the agrarian economy… now about 3%... It was the greatest century of an economy … in the history of civilization.”
— Rick Heitzmann, 37:35

“I have confidence in [Gen Z]… the best people are going to be the people who understand it a little ahead of time.”
— Rick Heitzmann, 40:58

8. Lightning Round: Startup Portfolio Deep Dive (49:12 - 56:49)

Discord (49:29)

• The shift from public, feed-driven social media to private group chats is a net positive for quality and engagement, though echo-chamber risks remain.

“Curation has been the most important thing to keeping a good, thriving community.”
— Rick Heitzmann, 51:41

DraftKings (51:54)

• Legal and regulatory frameworks lag behind technology; expect more rules on insider sports betting and market integrity.

Shopify (53:12)

• Not all commerce will move to chatbots—physical infrastructure and fulfillment remain essential; AI will change the front-end experience but not the back-end engine of commerce.

“There still needs to be a T-shirt in a warehouse … which needs to ship to you… the commerce infrastructure is not going to go away regardless of who initiates that transaction.”
— Rick Heitzmann, 53:50

Airbnb (56:00)

• New York City’s Airbnb ban is described as a self-inflicted wound that fails to address underlying housing policy problems.

Selected Notable Quotes

• “We haven't seen this wave of startups that we believe are sustainable. … They're not that step function different.”
  — Rick Heitzmann, 06:05

• “Is that data privacy going to be a key limiter to how the next generation of companies evolve?”
  — Rick Heitzmann, 13:18

• "Only the expert is going to sit on top and say, ‘Hey, I'm going to be your dietitian, I'm going to use the backend of ChatGPT like you would, but I'm going to give some more advice...’"
  — Rick Heitzmann, 17:03

• "The economy is very strong … The second piece of that is I do believe that companies are slow to hire. … Companies are now thinking, ‘How are we more efficient?’"
  — Rick Heitzmann, 47:21

Timestamps for Key Sections

• 00:33 — Where are the AI startups? The consumer AI drought.
• 02:17 — Why broad consumer AI is hard to disrupt; data as differentiator.
• 07:29 — Wrappers vs. native applications; defensibility challenges.
• 13:21 — Data privacy and enterprise opportunities.
• 16:03 — Human value-add and creative destruction.
• 20:24 — AI companionship, ethical concerns.
• 23:22 — The scale of AI funding and the investment boom.
• 31:50 — Dependence of public/private markets on OpenAI’s success.
• 36:58 — Will AI automate all work? Historical lessons and labor shifts.
• 41:28 — Corporate job "waste" and AI as a tool for efficiency.
• 49:12 — Discord, DraftKings, Shopify, Airbnb portfolio lightning round.

Conclusion

This episode presents a wide-ranging and nuanced look at why the explosion of AI-driven consumer startups has not materialized, the unique value of proprietary data in enterprise AI, and the profound economic implications of the generative AI revolution. Heitzmann offers cautious optimism for creative adaptation—as long as businesses and individuals keep finding and building on the unique value humans (still) bring to the table.