What will be OpenAI’s IPO price? Place ya bets! | E2202
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 Halloween episode of This Week in Startups, Jason Calacanis and Alex Wilhelm dive deep into the latest developments in AI, prediction markets, and startup dynamics. The show kicks off with Jason's experience using Tesla's new "Mad Max" mode in FSD 14.1, which he describes as noticeably more aggressive and better than previous versions (02:26). The hosts then explore Brian Armstrong's punk rock moment when he gamed prediction markets during Coinbase's earnings call by spontaneously mentioning key terms that people were betting on (05:08).
- Core themes include AI tool evolution, the dangers of platform dependency, prediction market manipulation, and the trust-based nature of the venture capital ecosystem
Speakers
Jason Calacanis
Jason Calacanis is a prominent angel investor, serial entrepreneur, and podcast host who has been active in Silicon Valley for over two decades. He's the founder of Launch, an investment company and accelerator, and has made early investments in companies like Uber, Robinhood, and Thumbtack. He's also the host of This Week in Startups and co-host of the All-In Podcast.
Alex Wilhelm
Alex Wilhelm is a seasoned technology journalist and newsletter writer who covers startups, venture capital, and public markets. He writes a daily newsletter focused on startup and tech analysis, providing insights into funding rounds, IPOs, and market trends. He brings a data-driven perspective to startup and venture capital discussions.
Key Takeaways
Avoid Building on Platforms That Want to Own Your Business
Jason issued a stark warning about using OpenAI's API, comparing Sam Altman to Mark Zuckerberg and Bill Gates in their platform strategies (50:15). He explained how these companies study API usage patterns to identify successful applications they can replicate internally. The context came from discussing how OpenAI is aggressively expanding into every vertical - from job search to travel to social media - while collecting data on which companies use their services most heavily. This creates an existential risk for any company building on their platform, as they're essentially providing a roadmap for OpenAI to compete directly with them.
Startups Are Creating Their Own AI Models to Reduce Dependency
Three major startups - Cursor, Canva, and Windsurf - all launched their own AI models this week, signaling a shift away from paying margins to major AI providers (46:50). These models are reportedly based on retrained Chinese open-source models and perform comparably to state-of-the-art options while being faster and cheaper. This trend represents startups taking control of their destiny rather than remaining dependent on OpenAI, Anthropic, or other major providers who might eventually compete with them directly.
The Music Industry Operates Like a Legal Trap for Startups
Jason shared his theory about how the music industry deliberately allows startups to use their IP initially, then strikes when the companies achieve success (17:03). He compared their approach to organized crime tactics, where they let you "gamble" with their content, build up debt, and then take over your business through coordinated lawsuits. His advice is to never touch music industry IP, only use work-for-hire musicians, and create royalty-free content from scratch. The discussion was prompted by Udio settling with Universal Music Group, showing how these partnerships typically involve significant compensation to the labels.
Trust-Based Venture Capital System Can Be Exploited
Jason revealed a concerning trend where some founders are keeping investor money without building products, exploiting the loose structure of SAFE notes (66:57). Unlike convertible notes with protective provisions, SAFE notes can go on indefinitely with minimal oversight, allowing bad actors to take advantage of the trust-based system. He emphasized that founders should play the long game, cherishing early investors who believed in them when nobody else would, rather than screwing them over when later investors offer better terms.
AI-Enhanced Search Will Drive Massive Growth in Query Volume
Contrary to fears that AI will cannibalize Google's search business, Jason argued that better AI-powered results will actually increase search volume by 10x (62:59). He explained that when people can get genuinely helpful answers - like diagnosing appliance problems or HVAC issues without calling repair services - they'll search more frequently. This insight came from analyzing Google's earnings, which showed search volume continuing to grow despite AI integration, supporting his thesis that improved product quality drives increased consumption.
Statistics & Facts
- Azure growth at Microsoft was 40% in the most recent quarter, Google Cloud grew 34%, and Amazon AWS grew 20.2% - the best performance for Amazon in 11 quarters (39:49). This demonstrates continued strong demand for cloud computing infrastructure despite concerns about AI spending.
- Google increased their CapEx estimate to $91-93 billion for this year, up from their previous estimate of $85 billion (40:11). This shows major tech companies are still compute constrained and willing to invest heavily in AI infrastructure.
- Suno reached $150 million in ARR according to Jason (16:58), demonstrating the massive market opportunity for AI-generated music tools that enable non-musicians to create professional-quality tracks.
