AI & us

AI & us

#04 Inside Claude Sonnet 4’s System Prompt

20 min · 28 de may de 2025
portada del episodio #04 Inside Claude Sonnet 4’s System Prompt

Descripción

AI & us – Episode #04 – Inside Claude’s Head: Secrets of the Sonnet Prompt This week on AI & us, we dive deep into the mind of Claude Sonnet 4 – literally. We got our hands on Claude’s secret system prompt, and we’re unpacking it all: *  Why Claude acts more like a thoughtful friend than a robot assistant *  How it coaches your prompts (without sounding like a jerk) *  The hardcoded rules that stop it from going full Skynet *  And yes… there’s poetry critique, CEO name-drops, and a strict no-lyrics policy It’s part AI philosophy, part digital etiquette, part employee handbook – and fully fascinating. Join us for the prompt behind the personality – and meet the Claude you didn’t see coming. This podcast is generated by AI

Comentarios

0

Sé la primera persona en comentar

¡Regístrate ahora y forma parte de la comunidad de AI & us!

Prueba gratis

Empieza 7 días de prueba

$99 / mes después de la prueba. · Cancela cuando quieras.

  • Podcasts solo en Podimo
  • 20 horas de audiolibros al mes
  • Podcast gratuitos

Todos los episodios

8 episodios

episode #07 Power Hungry: AI Goes Military, Nuclear & Viral artwork

#07 Power Hungry: AI Goes Military, Nuclear & Viral

🎧 AI & Us – Episode 7: Power Hungry: AI Goes Military, Nuclear & Viral Welcome to another rollercoaster ride through the world of artificial intelligence! This week, we’re diving into AI’s biggest appetites – for cash, kilowatts, and, well… your privacy. 👉 What’s inside: Frontier model, front-line budget: OpenAI grabs a $200M Pentagon contract for “non-kinetic” AI. We unpack what that means, why RAG and RLSF matter, and whether public AI is becoming the budget version of military-grade AI. Gen-Z homework hacks go VC: Berlin’s Knowunity raises €27M to build a TikTok-style AI tutor for a billion students. Is this the future of learning or a fast track to AI-fueled plagiarism? Watts for bots: Inside the closed-door ENACT “Power Summit,” where oil execs, AI labs and policymakers plot how to keep the GPU beast fed – from mini nuclear reactors to flare-gas data centers. ⚙️ Prompt Patrol: Microsoft’s Copilot Vision lets AI see your screen (with permission) and guide your apps – finally, Clippy gets laser eyes. 🤖 Bot or Not?: A Meta AI UX fail turns private therapy-level chats into public cringe in the Discover feed. We collect the jaw-droppers so you don’t have to. 👉 Subscribe now before an LLM does it for you. “Where tech gets real, and reality gets a reboot.” This podcast is generated by AI.

20 de jun de 202518 min
episode #06 AI Gets Real artwork

#06 AI Gets Real

🎙️ AI & us – 06 "AI Gets Real" 🔍 This week on AI & us: AI is no longer a lab experiment – it's showing up in your commute, your apps, and your OnlyFans feed (seriously). In this episode, we explore how artificial intelligence is stepping into the real world – subtly, powerfully, and sometimes weirdly. --- 🧠 Feed me up, Scotty! – This week’s top 3 AI stories: 1. 🚘 Tesla's Robotaxi Launch Elon Musk sets the date: starting June 22, driverless Tesla robotaxis will begin operating in Austin, Texas. ➤ Are we ready for a car with no steering wheel – and no human behind it? 2. 🍏 Apple Intelligence goes official At WWDC 2025, Apple revealed its AI play: “Liquid Glass” UI, smarter Siri, private on-device models, and more. ➤ Quiet revolution or too little, too late? 3. 🇪🇺 Nvidia & Perplexity’s European AI alliance Custom AI models for all 24 EU languages? That’s the goal of Nvidia and Perplexity’s new partnership – with open-source tools and sovereign infrastructure. ➤ A real step toward digital independence? --- ⚙️ Prompt Patrol – This week’s tool: NotebookLM by Google Think of it as a research assistant that actually reads your PDFs. ➤ We used it to prep this episode. Here’s why you might want it in your workflow too. --- 🤖 Bot or Not? – Things get weird: Katie Price resurrects “Jordan” as an AI avatar Her ‘90s alter ego is back – now in 3D, AI-powered and available 24/7 on OnlyFans. ➤ Genius marketing or uncanny valley nightmare? --- 🗨️ Feedback? Thoughts? Rants? Voice memos generated by your AI clone? Drop us a line on LinkedIn – we’d love to hear from you: https://www.linkedin.com/in/jan-hill-250695209/ [https://www.linkedin.com/in/jan-hill-250695209/]  This podcast is generated by AI.

13 de jun de 202516 min
episode #05 Pump the Brakes, AI? artwork

#05 Pump the Brakes, AI?

🎙️ AI & us – Episode 5: Pump the Brakes, AI? In this episode of AI & us, your favorite duo dives into the most jaw-dropping, frustrating, and fascinating stories from the world of artificial intelligence. 🧠 Feed me up, Scotty! * Anthropic’s CEO predicts AI will kill 50% of entry-level white-collar jobs. Is this a warning – or a wake-up call? * 77% of Americans say: Slow. It. Down. A new poll shows the public wants AI progress to hit the brakes – across all generations. * UK copyright battle heats up as creators push back against AI companies using their work without consent. Are we about to rewrite copyright law for the AI age? ⚙️ Prompt Patrol We spotlight KODISC – a classroom-ready AI tool built by a high school student (!) that turns your questions into full-blown explainer videos. 🤖 Bot or Not? A fantasy author forgets to delete a ChatGPT prompt from her novel. The result? Viral embarrassment, one-star reviews, and a bigger debate about AI in creative writing. 🎧 Whether you're coding, commuting, or just curious: this is your weekly upgrade to what really matters in AI. No hype, just the good stuff – with context, humor, and opinions that aren’t afraid to take sides. Subscribe now and get smarter, one episode at a time. → AI & us: Where tech gets real, and reality gets a reboot.

6 de jun de 202517 min
episode #04 Inside Claude Sonnet 4’s System Prompt artwork

#04 Inside Claude Sonnet 4’s System Prompt

AI & us – Episode #04 – Inside Claude’s Head: Secrets of the Sonnet Prompt This week on AI & us, we dive deep into the mind of Claude Sonnet 4 – literally. We got our hands on Claude’s secret system prompt, and we’re unpacking it all: *  Why Claude acts more like a thoughtful friend than a robot assistant *  How it coaches your prompts (without sounding like a jerk) *  The hardcoded rules that stop it from going full Skynet *  And yes… there’s poetry critique, CEO name-drops, and a strict no-lyrics policy It’s part AI philosophy, part digital etiquette, part employee handbook – and fully fascinating. Join us for the prompt behind the personality – and meet the Claude you didn’t see coming. This podcast is generated by AI

28 de may de 202520 min