Tech nostalgia. Winner-take-all economics. The cult of "storytelling". A Stanford educated aristocratic elite. This was the week that nothing changed in Silicon Valley. Alternatively, it was the week that radical change broke some ChatGPT users hearts. That, at least, is how That Was the Week tech newsletter publisher Keith Teare described this week in Silicon Valley. From Sam Altman's sensitivity to user backlash over GPT-5's personality changes, to venture capital's continued concentration in just ten mega-deals, to Geoffrey Hinton's apocalyptic warnings about AI wiping out humanity - the patterns remain stubbornly familiar even as the technology races forward. So is nothing or everything changing? Keith says everything, I say nothing. Maybe - as AI Godfather Hinton suggested on the show earlier this week - it's time for an all-knowing algorithm with maternal instincts to enlighten us with the (female) truth about our disruptive future.

1. AI Users Are Forming Deep Emotional Bonds

ChatGPT users experienced genuine heartbreak when GPT-5's personality changes made their AI feel like a different "person." This forced OpenAI to backtrack and restore GPT-4, revealing how humans are treating AI as companions rather than tools.

2. Silicon Valley's Power Structures Remain Unchanged

Despite AI's revolutionary potential, the same patterns persist: 40% of VC money goes to just 10 deals, Stanford maintains legacy admissions favoring the wealthy, and winner-take-all economics dominate. The technology changes; the power concentration doesn't.

3. The Browser Wars Are Over - Chat Interfaces Won

The future battle isn't about owning browsers (like Perplexity's bid for Chrome) but controlling the chat interface. OpenAI and Anthropic are positioning themselves as the new gatekeepers, replacing Google's search dominance.

4. AI's Pioneers Are Becoming Its Biggest Skeptics

Geoffrey Hinton, the "AI godfather," now believes there's a 15-20% chance AI could wipe out humanity. When the field's leading experts admit they "have no clue" about AI's future risks, it reveals how little anyone really knows about what we're building.

5. Context and Prompting Are the New Programming

The era of simple AI prompts is over. Success now requires sophisticated prompt engineering and providing rich context - making AI literacy as crucial as computer literacy once was. The abstractions are changing, and so must our skills.

Keen On America is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit keenon.substack.com/subscribe

Episode #221: Continuous Glucose Monitors

Barbell Medicine Podcast

Exploring Glucose Levels in Athletes and Upcoming Content Highlights

The AI-powered Podcast Player