Dan's AI Intel

Daniel Walter

The AI revolution is moving faster than anyone can follow. I'm Dan — I built my own stack of AI tools to research, verify and explain the developments I can't stop thinking about, and I share what I find as a two-host show you can actually keep up with. For curious people who want to understand where AI is really going, not just the headlines. AI-assisted, fact-checked, worth a second look.

  1. Claude Sonnet 5 Got Cheaper — So Why Does It Cost More?

    4d ago

    Claude Sonnet 5 Got Cheaper — So Why Does It Cost More?

    Claude Sonnet 5 is cheaper per token and nearly matches Opus 4.8 — so why does finishing a real task cost more than before, and more than Opus itself? We dig into the new tokenizer, the 3x jump in agent loops, how Sonnet 5 stacks up against Gemini 3.5 Flash on price and speed, and the one metric that actually decides your AI bill: cost per completed task, not price per token. I'm Dan. AI moves too fast to keep up with, so I built my own stack of AI tools to research, analyse, verify and illustrate the questions I can't stop thinking about — mostly to learn it myself, and I share what I find. AI-assisted, fact-checked, worth a second look. Follow Dan's AI Intel in whatever app you're listening in — it's free, and a follow is the single biggest thing that helps a small independent show grow. Read the full report — charts, sources and the interactive exhibits — at https://connectiveshift.com/shows/ai-intel/cc7df8a9-3357-49e1-8bce-0f32e584d50f Prefer to listen? This episode is on Apple Podcasts https://podcasts.apple.com/au/podcast/claude-sonnet-5-got-cheaper-so-why-does-it-cost-more/id6784551302?i=1000775767746 — or follow the show wherever you get your podcasts. One quick thing, from Dan — I make this show mostly to keep up with AI myself, and I'd love to make it better. If there's something you'd push back on or want me to go deeper on, tell me: podcast@connectiveshift.com — I read every one.

    30 min
  2. Dario Amodei Just Asked for a Kill Switch on His Own AI

    5d ago

    Dario Amodei Just Asked for a Kill Switch on His Own AI

    Dario Amodei just asked the U.S. government for the power to block an AI model from shipping — and the model he meant was his own. We trace Anthropic's CEO across his four big essays, from "Machines of Loving Grace" to "The Adolescence of Technology" to June's "Policy on the AI Exponential," unpack the five things he now wants Washington to do, and ask the hard question: when only the people racing to build AI understand it well enough to govern it, is a kill switch real safety — or the perfect moat? With Sam Altman, Yann LeCun and Geoffrey Hinton along the way. I'm Dan. AI moves too fast to keep up with, so I built my own stack of AI tools to research, analyse, verify and illustrate the questions I can't stop thinking about — mostly to learn it myself, and I share what I find. AI-assisted, fact-checked, worth a second look. Follow Dan's AI Intel in whatever app you're listening in — it's free, and a follow is the single biggest thing that helps a small independent show grow. Read the full report — charts, sources and the interactive exhibits — at https://connectiveshift.com/shows/ai-intel/d471e1e6-2b66-4da3-bf8a-a29da008cf6d Prefer to listen? This episode is on Apple Podcasts https://podcasts.apple.com/au/podcast/dario-amodei-just-asked-for-a-kill-switch-on-his-own-ai/id6784551302?i=1000775726903 — or follow the show wherever you get your podcasts. One quick thing, from Dan — I make this show mostly to keep up with AI myself, and I'd love to make it better. If there's something you'd push back on or want me to go deeper on, tell me: podcast@connectiveshift.com — I read every one.

    38 min
  3. Google's Quiet AI Win: The Hidden 70x Subsidy in Your $200 Plan

    Jun 26

    Google's Quiet AI Win: The Hidden 70x Subsidy in Your $200 Plan

    Everyone thinks AI's per-token meter is the subsidy. It isn't — inference is profitable. The real giveaway is your flat $200 ChatGPT Pro or Claude Max plan, a 40–70x subsidy to its heaviest users. And who survives the price war is decided at the silicon: Google runs Gemini on its own TPUs while OpenAI, Anthropic, xAI and the rest pay NVIDIA's ~75% margin on every token. We map every major maker, the reasoning-token trap that makes the cheap model the expensive one, why prices fall as losses hit records, and when the meter comes for your subscription. I'm Dan. AI moves too fast to keep up with, so I built my own stack of AI tools to research, analyse, verify and illustrate the questions I can't stop thinking about — mostly to learn it myself, and I share what I find. AI-assisted, fact-checked, worth a second look. Follow Dan's AI Intel in whatever app you're listening in — it's free, and a follow is the single biggest thing that helps a small independent show grow. Read the full report — charts, sources and the interactive exhibits — at https://connectiveshift.com/shows/ai-intel/ad0f6c8b-7a75-4dd8-b288-78d3fba4dc9d Prefer to listen? This episode is on Apple Podcasts https://podcasts.apple.com/au/podcast/googles-quiet-ai-win-the-hidden-70x-subsidy-in-your-%24200-plan/id6784551302?i=1000774287939 — or follow the show wherever you get your podcasts. One quick thing, from Dan — I make this show mostly to keep up with AI myself, and I'd love to make it better. If there's something you'd push back on or want me to go deeper on, tell me: podcast@connectiveshift.com — I read every one.

    43 min

About

The AI revolution is moving faster than anyone can follow. I'm Dan — I built my own stack of AI tools to research, verify and explain the developments I can't stop thinking about, and I share what I find as a two-host show you can actually keep up with. For curious people who want to understand where AI is really going, not just the headlines. AI-assisted, fact-checked, worth a second look.