Episode Summary
Executive Summary: The conversation explores how AI is reshaping the value of data, why community may become more important in a digital world, and how the guest thinks about founder talent, succession, and decision-making. He argues that careers and companies are best built around short-term enjoyment, fast learning, and niche communities, while warning against over-optimizing for resume status or midterm goals.
Main Topics: Data as an AI-era commodity (Priority: 5/5): The guest revisits an earlier thesis that data would become one of the world’s most valuable resources, then reassesses it in light of current AI buyers like OpenAI, Anthropic, Google, and others. He is skeptical that data demand has scaled as much as expected outside a small set of buyers. Community as a durable human need (Priority: 5/5): He argues that in an AI-saturated world, real human community—especially repeated in-person interaction—will become even more valuable. He sees community as both personally meaningful and strategically important, even if it is not always the best business model. Founder archetypes, talent, and succession (Priority: 5/5): The guest discusses whether successful founders share an archetype, concluding that successful people are highly diverse. He emphasizes finding exceptional people, giving them room to grow, and promoting from within when possible rather than relying on external executive hires. Decision-making across time horizons (Priority: 4/5): He lays out a framework for micro-time, short-time, long-time, and big-time decision-makers. He places himself in the short-time bucket and argues people should know their strengths, especially how quickly they can make good decisions. Who you know vs. what you know (Priority: 4/5): He argues that ‘what you know’ became more important in the 2000s as knowledge became easier to discover, but who you know still matters socially and professionally. In AI, he says the balance could shift again, but he remains uncertain. Insiders, outsiders, and freedom (Priority: 4/5): He reflects on the insider/outsider spectrum in elite circles, citing examples such as Mark Andreessen, Peter Thiel, and Paul Graham. His view is that insider status brings benefits but reduces freedom, and some people intentionally move back toward outsider status. Books, TV, and learning preferences (Priority: 2/5): He recommends concise, high-signal nonfiction and specific works like Zero to One and Difficult Conversations, while criticizing the habit of reading too many books. He also shares a list of favorite shows, emphasizing high-quality storytelling and rewatchability.
Key Arguments: Data became more valuable in AI, but the market for raw data is still narrower and less recurring than expected; only a few AI labs and companies are major buyers. Community will matter even more in the future because AI cannot fully replace deep, repeated, in-person human connection. Successful founders do not share one fixed archetype; persistence may be common, but most traits vary widely by person. To help exceptional talent grow, leaders should increase responsibility and allow a meaningful chance of failure rather than over-scaffolding. Career decisions should optimize for short-term enjoyment and long-term direction, not the midterm or resume value. People who spend the most time talking in a meeting usually learn the least; learning often comes from engaging with people a step or two below you. Talent evaluation is extremely hard; interviews and resumes help filter, but they rarely identify the true 10X person in advance. People change over time, so old assessments, roles, and ambitions should be periodically reevaluated. AI may change the relative value of ‘who you know’ and ‘what you know,’ but the guest still sees ‘what you know’ as more important today. Many popular ideas and movements are aided by implicit coordination or back-channel incentives, even when they are not purely conspiratorial.
Data Points: OpenAI data spending: a lot - Cited as one of the major AI buyers of data today. Anthropic data spending: a lot - Used as an example of AI labs buying substantial amounts of data. Google Reddit data deal: large spending - Referenced as a major purchase of data by Google. OpenAI–New York Times deal: $20 million - Mentioned as possibly being more about press/public relations than raw data value. World of Das YouTube subscribers: 70,000 - The podcast’s channel subscriber count was cited as evidence of growth. Data-buying hedge funds at peak: about 100 - Estimated number of serious hedge-fund data buyers when SafeGraph was founded. Data-buying hedge funds today: about 60 - Estimate given for current serious hedge-fund data buyers, showing decline. Typical failure-rate target for growth: one-third to two-thirds - He argues this is the ideal range of failure for people optimizing for growth. Very high failure rate: 98% - Used as an example of too much failure, which would be counterproductive. Flex Capital seed pace: about 80 deals/year - The fund’s current annual pace of seed investments. Flex Capital target volume: 100 checks/year - Stated as the goal for the seed investing strategy. Typical Flex check size: $300K–$500K - Approximate check size for seed investments. Church/community size example: 30,000+ people - Used to illustrate how large institutions can lose intimate community. Small church/community size example: 200 people - Used as the scale where everyone can know each other.
Pivotal Quotes: "The one thing I'm more confident on in the world of AI is that things like community of people is going to become even more important." — Aarne: He explains his strongest view about the future in AI is the rise of real human community. "I think it is still better to be a what you know than a who you know." — Aarne: His current view on career advantage in the AI era, while acknowledging uncertainty over the next decade. "I would say somewhere between one third and two thirds is like the right ... failure rate is like the optimal point." — Aarne: His framework for how much failure ambitious people should experience to grow quickly.
Implications: Listeners should expect AI to increase the premium on authentic community, niche identity, and real expertise. For founders and operators, the message is to favor growth, autonomy, and honest self-assessment over status-driven, midterm optimization.
About We the Builders
Conversations with practitioners at the edge of their craft across business, media, startups, frontier technologies, investing.