Jason Unpacks Sequoia’s New Funds"| E2199
Timestamps are as accurate as they can be but may be slightly off. We encourage you to listen to the full context.
Podcast Summary
This episode of This Week in Startups, hosted by Jason Calacanis and Alex Wilhelm on October 27th, covered a wide range of startup and tech industry topics. The hosts discussed the importance of upcoming earnings reports from Meta, Alphabet, Microsoft, and Amazon for AI infrastructure insights, examined fascinating drone show technology from China versus the West, and analyzed Tesla's breakthrough in using synthetic data for self-driving training. They also dove into Sequoia Capital's new fund strategy, the ethics of dark patterns in tech (particularly Microsoft's controversial subscription tactics), and the massive growth of AI training company Mercor. (03:44)
- Main themes: The episode explored AI infrastructure investment, technological competition between East and West, the importance of ethical business practices, and the rapid evolution of AI-assisted services across multiple industries.
Speakers
Jason Calacanis
Jason Calacanis is a renowned angel investor, entrepreneur, and host of This Week in Startups podcast. He's been investing in startups for over two decades and has backed companies like Uber (as a Sequoia Scout), Thumbtack, and numerous others. He's also the founder of Launch accelerator and has been podcasting for over 2,300 episodes, establishing himself as a leading voice in the startup ecosystem.
Alex Wilhelm
Alex Wilhelm is a technology journalist and co-host of This Week in Startups. He previously worked at TechCrunch where he covered startup funding, IPOs, and technology trends. He brings deep analytical expertise to the show, particularly around funding dynamics, market analysis, and startup metrics.
Key Takeaways
Pay Attention to Big Tech Earnings for AI Infrastructure Insights
Jason emphasized that startup founders should unusually pay attention to this week's earnings from Meta, Alphabet, Microsoft, and Amazon - companies he calls the "hyperscaler set." (14:23) These companies have been pouring tens of billions into AI infrastructure and previously reported being "compute constrained" with insufficient GPU access. Their reports will reveal not just trailing performance but also forward-looking guidance about AI demand, industry adoption, and infrastructure capacity. This data directly impacts the overall AI opportunity landscape that startups operate within, making it unusually relevant for founders to track.
Avoid Dark Patterns - They Damage Long-term Success
The discussion of Microsoft's controversial subscription tactics in Australia led to a crucial lesson about business ethics. (69:07) Jason stressed that dark patterns - deceptive UX designed to trick users - are particularly damaging when you're winning. He referenced a $2.5 billion FTC settlement against Amazon for similar practices, emphasizing that large successful companies don't need these tactics and they build resentment over time. For startups under pressure to grow, Jason recommended the "mom test" - if your mother wouldn't approve of the practice, don't do it.
Structure Incentives for Long-term Customer Success
When discussing how startups can avoid falling into dark pattern traps, Jason provided a framework for aligning team incentives with customer success. (77:39) Instead of paying sales teams only for initial bookings, he suggested splitting compensation: lower initial commission but additional bonuses tied to customer usage metrics, renewal rates, and long-term success. For example, paying 7% upfront, then 5% more if customers actively use the product, and additional bonuses for renewals. This creates alignment between team success and customer satisfaction.
Synthetic Data Will Accelerate Self-Driving Development
Tesla's demonstration of fully synthetic training data for self-driving cars represents a breakthrough in AI development methodology. (28:56) The ability to generate unlimited diverse driving scenarios - from deer running into streets to complex weather conditions - allows for much more comprehensive training than waiting to capture rare real-world events. This approach, also being pursued by competitors like Waymo and Uber's partnership with Nvidia, suggests that synthetic data generation will become a critical competitive advantage across AI applications, not just autonomous vehicles.
Geographic and Cultural Knowledge Matter for Investment Success
When discussing Sequoia's international investment strategy, Jason shared his personal investment philosophy that directly applies to startup expansion decisions. (44:55) He avoids investing in regions where he doesn't speak the language, understand the culture, or know the legal framework, believing local investors better serve those entrepreneurs. This principle extends to startups considering international expansion - success requires deep local knowledge, not just product-market fit. Founders should partner with local experts or ensure they have team members who truly understand target markets before expanding globally.
Statistics & Facts
- Amazon employs 1.4 million people, which represents approximately 1% of the entire US workforce, making it one of the largest private employers in the country. (01:35) This statistic was mentioned in the context of discussing Amazon's shift toward robotics and automation.
- Only 3% of journalists now identify as Republican/conservative, compared to a more balanced 27% Republican, 35% Democrat, and 40% Independent distribution in the 1970s-80s. (41:20) Jason used this to explain bias in media coverage and the importance of diverse journalistic perspectives.
- Waymo has driven over 95 million miles across their markets (16M in LA, 30M in San Francisco, 46M in Phoenix, 3M elsewhere) with significantly improved safety metrics: 91% fewer serious injuries, 79% fewer airbag deployments, and 92% fewer pedestrian crashes with injury compared to human drivers. (39:19)
