Slow Hunch Podcast on Emergence

[Alex Komoroske]

A podcast I recorded with Nick Grossman from Union Square Ventures was just released:

https://www.youtube.com/watch?v=c6fBc0vG5zE

The main topic was emergence, but here are a few other topics, courtesy of Claude:

• "Transformers were inevitable, not magical" - Why AI's emergence follows the same evolutionary patterns as the prokaryote-to-eukaryote transition, making LLMs a biological inevitability rather than a lucky discovery
• The software economics flip nobody's talking about - How AI inverts 40 years of assumptions: software is now free to write but expensive to run, breaking every business model we know
• OpenAI's reckless security gamble - Why prompt injection is fundamentally unsolvable if LLMs are "calling the shots," and why OpenAI's recent browser integration is "shooting fish in a barrel" for attackers
• The AOL parallel - Why OpenAI's dominance mirrors AOL's rise (and likely fall), and how ChatGPT's brand power masks structural weaknesses
• Wikipedia's hidden lesson for AI - How "folksonomies" and emergent systems from Web 2.0 show us what we're missing in current AI architectures
• The "infinite software" paradigm - Why apps are the wrong distribution model for AI, and how we need to flip from "apps own data" to "data carries its own policies"
• Confidential compute as the missing key - How military-grade encrypted cloud computing could enable privacy-preserving AI that doesn't require trusting any company
• Why vibe coding can't scale - The fundamental trust problem when random people can build apps that access your data