E1013: Decade Wrap-Up Roundtable! Garry Tan, Initialized Capital & Zach Coelius, Coelius Capital on the biggest hits & misses of the past decade: Trump, Theranos, Streaming, Mindfulness, Crypto, Tesla, WeWork, Amazon Prime, Vaping & more!

Episode Summary

Executive Summary: Jason Calacanis hosts Gary Tan and Zach Coelius for a decade-in-review bracket debate on the biggest startup trends and worst failures of the 2010s. They crown AI/ML, ride-sharing, streaming, and mindfulness as defining wins, while election manipulation, Facebook/Zuckerberg, WeWork, Theranos, and related hype cycles represent the darkest failures. The conversation ties startup innovation to societal shifts in labor, media, governance, and mental health.

Main Topics: AI/ML as the defining technology trend (Priority: 5/5): Both guests argue machine learning and AI were the most important startup trend of the decade, pointing to rapid gains in computer vision, speech recognition, and real-world deployments like drive-thrus and cashierless retail. Gig economy and ride-sharing (Priority: 5/5): Ride-sharing is framed as a decade-defining business model that reshaped labor, pricing, and consumer behavior, with broader implications for flexibility and quasi-UBI via on-demand work. Mindfulness vs. streaming as cultural winners (Priority: 4/5): The bracket turns philosophical, contrasting attention capture and entertainment streaming with the rise of mindfulness as a response to anxiety, social media pressure, and emotional instability. Platform power, governance, and social media harm (Priority: 5/5): Zuckerberg/Facebook, political ads, and the erosion of truth are discussed as major negative forces, with concerns about speech control, disinformation, and the scale of tech companies resembling governments. Fraud, hype, and governance failures (Priority: 5/5): WeWork, Theranos, and the ICO boom are used as cautionary tales of weak oversight, inflated narratives, and abuse of investor enthusiasm. Future risks: climate, nuclear, China, and biotech (Priority: 4/5): The discussion closes by projecting the 2020s around carbon pricing, nuclear power, China’s political trajectory, and gene editing/eugenics as major macro forces.

Key Arguments: AI/ML won the decade because deep learning crossed from promise to practical deployment in products like computer vision and voice systems. The gig economy boosted labor flexibility, helped keep unemployment low, and gave workers more control over hours and income. Streaming is culturally powerful because it provides abundant, high-quality entertainment and can shape society through narrative and escapism. Mindfulness is important because the decade was marked by anxiety, social-media-driven instability, and a loss of emotional self-control. Facebook/Zuckerberg represents a dangerous concentration of power because scale, copied ideas, and weak governance can distort speech and democracy. Election misinformation and microtargeted political ads may have damaged democratic truth structures and could have long-term civilizational consequences. Theranos and ICOs exemplify fraud and hype, but Theranos is viewed as the more deranged and sustained deception. WeWork is criticized primarily as a governance disaster rather than a product disaster, showing the need for stronger board oversight. Carbon pricing and climate action are presented as urgent because unpriced emissions are a massive economic and environmental externality. Nuclear energy, CRISPR, and possible political change in China are highlighted as likely major forces in the next decade.

Data Points: YC fund size: $7 million initial fund in 2012; now fourth fund at $225 million - Gary Tan describes the growth of Initialized Capital Syndicate size: 3,300 members - Zach Coelius explains the scale of his AngelList syndicate Diligence network size: 1,800 people - Zach says his syndicate acts like a huge scout and diligence network Lower unemployment attribution: Gig economy helps keep unemployment low - Calacanis links ride-sharing and on-demand labor to labor-market flexibility Drive-thru AI exit: Apprente sold to McDonald's - Zach cites the sale as proof that AI is already producing large exits Computer vision improvement: Dog-vs-cat error rate fell from about 30% to 2%-3% - Gary uses ImageNet-style progress to illustrate rapid AI progress Amazon Prime price: $150 per year - Calacanis mentions the subscription cost while describing its behavioral lock-in Tesla Model 3: Driven to the studio by Jason - Used as evidence of Tesla’s product and self-driving progress Seed check range: $1 million to $5 million - Gary describes Initialized’s typical entry stage Alternate seed check range: $200k to a couple million - Zach describes the stage range he likes for early investing Carbon pricing estimate: $40 per ton - Jason argues emissions should be priced to address climate change Economic scale of carbon pricing: $40 trillion per year - Jason estimates the implied value of unpriced carbon emissions Political ad cost gap: Trump about $1 vs. Clinton about $10 per ad - Jason describes how better engagement optimization lowered Trump’s ad costs

Pivotal Quotes: "I think that we're still at the early days of what's actually happening there." — Gary Tan: On AI/ML being the biggest startup trend of the decade and still unfolding "It is a universal basic income. It is like a gig economy because you can pick how many hours you want to work." — Jason Calacanis: On ride-sharing and on-demand labor as a structural shift in work "I believe that mindfulness will conquer." — Jason Calacanis: Final vote declaring the cultural winner of the decade

Implications: The episode frames the 2010s as the decade of platform scale and tech-driven disruption, but also of backlash, fraud, and attention collapse. For listeners, the takeaway is that the 2020s will be shaped by regulation, AI, climate action, and better governance.