First interview with Scale AI’s CEO: $14B Meta deal, what’s working in enterprise AI, and what frontier labs are building next | Jason Droege
Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
Podcast Summary
In this episode, Lenny interviews Jason Droege, the new CEO of Scale AI, marking his first interview since taking over from Alex Wang after Scale's Meta deal. (01:32) Jason shares his journey from co-founding a company with Travis Kalanick before Uber to launching Uber Eats, which grew from zero to $20 billion in four and a half years. The conversation explores how AI models are actually getting smarter today - through expert humans teaching them specific skills rather than just ingesting more data. (13:33)
• The main theme revolves around the evolution of AI training from basic data labeling to sophisticated expert-driven model improvement, alongside practical entrepreneurship lessons from building massive consumer products
Speakers
Jason Droege
Jason is the new CEO of Scale AI, having taken over after the Meta deal where Alex Wang moved to lead Meta's super intelligence team. Before Scale, Jason co-founded a company with Travis Kalanick (called Scour) before Kalanick started Uber, and most notably launched and led Uber Eats from an idea to a $20 billion business that essentially saved Uber during the pandemic. (01:32) He has over 25 years of experience in technology and entrepreneurship, with a track record of building massive consumer-facing products and marketplace businesses.
Key Takeaways
Everything in Business is Negotiable
Jason learned this crucial lesson at age 19-20 when co-founding Scour with Travis Kalanick. During their funding process with investors like Ron Burkle and Mike Ovitz, deal terms constantly changed - from wanting 50% to 75% to 80% of the company, with ultimatums and breakfast meetings as leverage tactics. (06:47) This taught him that there's no "standard way" to do business - if you can imagine it, align incentives, and negotiate effectively, almost anything is possible. This mindset later influenced both his and Travis's approach at Uber, where they constantly pushed boundaries and created non-traditional business arrangements.
Study Customer Incentives, Not Just Their Words
When launching Uber Eats, Jason couldn't get restaurateurs to share their unit economics directly, so his team ordered food and reverse-engineered the costs by weighing ingredients and matching them to supplier catalogs. (27:52) This revealed that restaurants pay roughly 20-30% for ingredients, 20-30% for labor, and 10% for real estate. Understanding that incremental demand could provide 70-80% gross margins on additional orders gave them confidence to charge 30% commission. The key insight: don't just listen to what customers say they need - understand their underlying business incentives and solve for those instead.
Focus on Problems with Daily Urgency
Jason emphasizes that customers might acknowledge a problem provides value, but if it's not something they think about urgently in their busy days, you'll have "a long road to a small town." (48:18) Using the restaurant example, while saving 3% on ingredient costs might be valuable annually, restaurateurs think daily about whether people showed up and whether they made money yesterday. Successful products solve problems that customers actively worry about, not just problems that objectively exist.
Survival is a Precursor to Winning
In a hype-driven tech culture that encourages "just go for it" mentality, Jason advocates for calculated risk-taking focused on not losing before trying to win. (65:05) He believes the best entrepreneurs make asymmetrically positive decisions throughout their journey and understand that survival through difficult periods is essential. Most founders give up before getting their timing right, finding the right customer insight, or achieving product-market fit. Life in tech can change quickly - you can go from struggling to successful rapidly, but only if you survive long enough for that transition to occur.
High Gross Margins Signal True Value Creation
Jason uses gross margins as a "coarse instrument" to quickly evaluate whether a business idea is differentiated and adding real value. (60:36) If you can't maintain high gross margins, it often means customers have good alternatives and you're not adding enough unique value. He suggests starting with the question: "Why can't we charge a 60% gross margin?" This forces you to identify the real competitive dynamics and understand whether your advantage will erode over time as others enter the market.
Statistics & Facts
- Scale AI has about 1,100 employees and operates two major business units, each generating hundreds of millions in revenue, essentially having "two unicorns inside the company." (12:12)
- Eighty percent of Scale's expert network contributors have a bachelor's degree or higher, with about fifteen percent holding PhDs. (16:00) This directly contradicts market narratives about the industry relying on low-skilled labor.
- Uber Eats grew from zero to $20 billion in revenue in just four and a half years, then jumped from $20 billion to $50 billion during COVID in approximately one year. (56:22) The business now generates around $80 billion in gross bookings.
