Better Offline Monologue: The Questions I'm Asking
73 snips
Dec 12, 2025 In a thought-provoking monologue, Ed Zitron dives into critical questions surrounding AI GPU warehousing and operational costs. He seeks insights on Blackwell GPUs, their failure rates, and actual revenue data from major AI companies. The discussion highlights the importance of revealing true AI operating costs to challenge existing industry perceptions. Ed's quest for accurate data aims to reshape funding in AI startups and improve transparency in the tech landscape.
AI Snips
Chapters
Transcript
Episode notes
Deals And Awards Are Mostly Theater
- Ed Zitron dismisses high-profile AI deals and awards as noise that won't change the industry's trajectory.
- He calls the Disney-OpenAI investment and TIME's AI person-of-the-year pick examples of hollow, performative industry moves.
Share GPU Warehouse Evidence
- If you have first-hand knowledge about NVIDIA AI GPUs being warehoused, contact Ed Zitron on Signal with pictures and numbers.
- Provide model types, ownership details, and cost information to materially improve reporting on GPU availability and costs.
Detail Blackwell GPU Costs And Failures
- Report any information on NVIDIA Blackwell GPU failure rates and operating costs, including hourly running costs.
- Ed Zitron emphasizes costs per GPU and reliability as key data points to understand the economics of AI compute.