AI market jitters, post-truth reality, data, and safeguarding what makes us human

This week we're taking stock of conversation trends to let it rip on AI market jitters and what happens when the math stops math-ing.We start with the numbers that have investors nervy: Amazon's $200 billion capex projection for 2026, and the uncomfortable reality of building an entire economy on depreciating GPU infrastructure with a three-year shelf life. Why the dot-com bubble comparison are incomplete, and questioning what happens when billions flow into overwhelming into transformer model architecture while research into others starves.Then we shift from market corrections to attention economics, unpacking how AI tools promise productivity while actually training us to outsource thinking itself. The cost is both financial and experiential. When was the last time you sat alone without reaching for your phone? Can you still read sentences that run four lines long?The episode lands on an uncomfortable question about who gets to have unmediated experiences anymore, and whether we're living our own lives or just consuming other people's.Mentioned: Ed Zitron ’s “Better Offline” podcast Derek Thompson’s Plain English podcast interview with Paul Kedrosky on market conditions and signs of a bubble Stephen Colbert on “truthiness” Enshittification, coined by Cory Doctorow MIT on the philosophical puzzle of AI Netflix’s main competition is sleep Point of view: Gen Z will remember more of other people’s memories than their own Blaise Pascal writing about attention in 1670