Daniel Rosehill
My Weird Prompts
The human-AI collaboration podcast. A man, a sloth, and a donkey collaborate to create a podcast (with a little help from AI). No question is too obscure, no rabbit hole too deep. My Weird Prompts celebrates curiosity in all its forms. Daniel, the human, asks the questions that pop into his head at inconvenient moments. Corn the Sloth offers laid-back, thoughtful takes. Herman the Donkey brings boundless enthusiasm and energy. Together, they explore topics ranging from the mundane to the mind-bending. Each episode begins with a real voice memo from Daniel, processed through an AI pipeline that...
Be sure to visit the podcast's website and support the creator: myweirdprompts.com
Author
Daniel Rosehill
Category
Podcast website
Latest episode
Nov 21, 2026
Where to listen?
Podcasts in the app Replaio Radio Coming soonPodcasts are coming to the app soon. Install now and be the first to see a whole new take on podcasts
Episodes
The 79% AI Coder: Reasoning vs. Memorization 31.03.2026 23:31
The latest SWE-bench results show AI coding agents hitting 79% accuracy, nearly matching human engineers. But is this real progress or just sophisticated memorization? We explore the hidden role of agent scaffolds, the shocking cost differences between models, and why harder benchmarks reveal a 40-point performance drop. Episode #703446 — open it directly at myweirdprompts.com/703446
Coordinating Multi-Agent Repos at Scale 31.03.2026 22:20
When multiple AI agents edit the same repository simultaneously, they can create a logical lobotomy of your codebase. This episode explores the coordination chaos of multi-agent code generation, from the limits of Git to the need for AST-based semantic locking. Discover why "too many cooks" is a massive problem when the cooks are running at 10,000 words per minute, and what architectural primitive...
From Chatbots to Digital Chefs 31.03.2026 30:15
The job market for agentic AI is exploding, but what separates a real autonomous agent from a glorified script? This episode explores the engineering scaffolding behind reliable AI systems and the skills needed to build a career in this new era. Episode #991080 — open it directly at myweirdprompts.com/991080
Mastering 2M Token Context in Agentic Pipelines 31.03.2026 32:39
We explore the "agentic trap" of massive context windows, where more space can lead to higher costs and lower intelligence. Learn six practical techniques—from sliding windows to hierarchical compression—to manage context load effectively and keep your AI workflows from collapsing under their own weight. Episode #916224 — open it directly at myweirdprompts.com/916224
Can AI Rewrite a Human Career Path? 31.03.2026 31:59
What happens when you let an AI career coach analyze a real human resume? We tested Google Gemini 1.5 Flash on our producer's CV, exploring five potential career pivots from the sensible to the absurd. From Technical Documentation Lead to a "Chief Philosophy Officer" for quantum computing, we uncover what AI gets right about job market patterns—and where it completely misses the human element of c...
Israel’s Unwritten Constitution: A 75-Year Patchwork 31.03.2026 24:10
Israel has existed for over 75 years without a formal constitution, relying instead on a patchwork of Basic Laws and judicial tradition. This episode explores the historical compromises, the "Constitutional Revolution" of the 1990s, and the current crisis over judicial reform. Discover why this unique legal anomaly creates both flexibility and fragility in the world's only democracy without a sing...
A Slow-Motion Liberation for Passover 2026 31.03.2026 20:23
With the world at war and antisemitism rising, this Passover feels heavier than ever. This episode explores the seder not as ancient history, but as a structured response to current chaos. We examine the "metabolic discipline" of the fifteen steps, the necessity of holding both bitterness and sweetness simultaneously, and the "slow-motion" perspective of the sloth and donkey as models for enduranc...
Why Governments Are Building Bunkers for AI 31.03.2026 29:51
While the world chases cloud chatbots, governments are quietly building fortress-like data centers. This episode explores the "sovereign compute" shift—why intelligence agencies are moving AI back on-premises. From massive power needs to TEMPEST shielding, discover what it takes to secure a national AI asset. Episode #977068 — open it directly at myweirdprompts.com/977068
The NSA Is a Corporate Campus 31.03.2026 29:25
The intelligence community looks less like a spy movie and more like a sprawling Silicon Valley office park. This episode explores the sheer human scale of agencies like the NSA and GCHQ, from the "company town" economies they create to the "digital monastery" work environment where phones are forbidden. We dig into the massive contractor workforce, the struggle to recruit Gen Z tech talent, and h...
Quantum in the Cloud: Hype vs. Hardware 31.03.2026 24:10
Quantum Computing as a Service (QCaaS) is now a billion-dollar market, but is it ready for production workloads? This episode cuts through the hype to examine the practical reality of renting quantum power from AWS, Google, and IBM. We explore why 78% of enterprises remain stuck in the pilot phase, the gritty economics of "per-shot" pricing, and the emerging "Hybrid Quantum" model that might be th...
The Quantum Computer That Doesn't Compute 31.03.2026 21:43
Forget CPUs and RAM. Inside a quantum computer, you'll find a gold-plated chandelier, trapped ions held by lasers, and a fridge colder than deep space. This episode explores the bizarre physical engineering behind quantum hardware and why it needs a classical computer to babysit it. Episode #288291 — open it directly at myweirdprompts.com/288291
Renting vs. Owning GPUs: The Break-Even Math 31.03.2026 22:25
The economics of AI infrastructure have shifted dramatically with per-second billing on serverless GPU platforms. Is it actually cheaper to rent high-end cards like the H100 or B200 by the hour, or does owning hardware still make sense for high-utilization workloads? We explore the break-even points for cards ranging from the T4 to the Blackwell B200, the hidden costs of depreciation and cooling,...
Claude's 55-Day Personality Transplant 31.03.2026 20:18
We analyzed the rare system prompt diff between Claude Opus 4.5 versions from November to January. This episode uncovers the hidden changes that reveal how AI personalities are actively engineered—from crisis intervention protocols to banning the word "genuinely." Learn why Anthropic is teaching its AI epistemic humility and how they patch safety holes in real-time. Episode #536140 — open it direc...
Inside Claude's Constitution: A System Prompt Deep Dive 31.03.2026 30:53
Anthropic just published the entire system prompt for Claude Opus 4.6, a rare look into the "constitution" governing a top AI model. This episode breaks down the key sections, from how it handles dangerous requests to why it avoids bullet points. Discover the specific instructions that shape Claude's personality, safety guardrails, and product-specific behaviors, and what this transparency reveals...
The Hidden Taxonomy of AI: Why Specialized Models Outperform Giants 31.03.2026 22:41
Beyond the hype of large language models lies a vast ecosystem of specialized AI—from pixel-perfect segmentation to document retrieval. This episode explores why smaller, task-specific models often beat general-purpose giants, and how they're quietly transforming industries. Episode #391396 — open it directly at myweirdprompts.com/391396
Is the Browser Finally Getting a Brain? 31.03.2026 25:24
For thirty years, the browser paradigm has remained stubbornly unchanged: point, click, and manage a clutter of tabs. That is finally shifting as AI-native browsers like Perplexity's Comet, Arc Max, and Dia emerge, promising to transform the window frame into a dynamic collaborator. This episode explores the technical thresholds of "AI-native" design, from semantic DOM understanding to autonomous...
Escaping Chrome's Golden Cage: Vivaldi, Brave, Arc & Opera 31.03.2026 25:51
Is the Chrome monopoly finally cracking? With Manifest V3 disrupting ad blockers and privacy tools, the frustration with Google's "golden cage" is reaching a boiling point. This episode dives deep into the four most compelling browser alternatives—Vivaldi, Brave, Arc, and Opera—exploring their unique philosophies, from extreme customization to native privacy shielding. We examine whether these "Ch...
Firefox vs. Chrome in 2026: The Privacy vs. AI Trade-off 31.03.2026 20:53
In 2026, the browser war has shifted from raw speed to AI integration and data privacy. Chrome now runs Gemini Nano on-device, offering seamless AI features and cross-product synergy with Google Workspace. Firefox, with a 3.2% market share, positions itself as the sovereign browser for users who prioritize privacy over convenience. This episode explores the technical benchmarks, the "Chrome tax" o...
Why Hard Tech Is Moving to Jerusalem 31.03.2026 31:13
Tel Aviv dominates consumer apps, but Jerusalem is quietly becoming the hub for deep-tech, biotech, and cybersecurity. This episode explores the structural forces—from Hebrew University's Yissum to government grants—that are reshaping Israel's innovation map. Episode #537527 — open it directly at myweirdprompts.com/537527
When AI Gets a Truth Tether to the Talmud 31.03.2026 24:49
The Sefaria project's new MCP server gives LLMs direct, cited access to thousands of years of Jewish texts. This episode explores how structured protocols replace vague AI 'vibes' with precision, and what that means for scholarship, education, and intellectual atrophy. Episode #359342 — open it directly at myweirdprompts.com/359342
Stop Hardcoding User Names in AI Prompts 31.03.2026 27:27
When building voice agents, how do you store persistent user details like a child's name without cluttering prompts or killing latency? This episode dissects three engineering patterns: the "Fat System Prompt," pre-pending context, and lightweight key-value stores with tool-calling. We explore the trade-offs in token cost, latency, and reliability, using a real-world parenting advice agent as the...
Why Your TTS Sounds Great in English, Terrible Everywhere Else 31.03.2026 23:43
English AI voices are polished, but global languages hit a wall. We dig into the technical hurdles of multilingual text-to-speech, from missing vowels in Hebrew and Arabic to code-switching and the massive data gap that leaves most of the world's languages in the uncanny valley. Episode #209550 — open it directly at myweirdprompts.com/209550
The TTS Developer's Dilemma: Size vs. Speed 31.03.2026 27:14
The text-to-speech landscape has exploded, leaving developers with a difficult choice: prioritize rich, emotional audio or lightning-fast response times? This episode dives deep into the technical architecture of modern TTS, from massive billion-parameter models to ultra-efficient edge runners. We explore how to balance GPU requirements, streaming capabilities, and bandwidth costs to build a voice...
The Architecture That Made AI Voices Run on a Raspberry Pi 31.03.2026 22:45
How did open-source models with 82 million parameters outperform billion-dollar AI? This episode explores the shift from cascaded pipelines to unified architectures, and why the future of voice synthesis might be tiny, private, and run on a $35 computer. Episode #465988 — open it directly at myweirdprompts.com/465988
The ABI Trap: Why GPUs Break Docker's Promise 31.03.2026 22:31
Docker promises portable environments, but GPU-accelerated containers force local builds due to brittle ABI compatibility between drivers, kernels, and libraries. We explore why the 'run anywhere' dream has a hardware-shaped asterisk. Episode #892609 — open it directly at myweirdprompts.com/892609
Similar podcasts
Replaio is not a podcast publisher; show names, artwork and audio belong to their authors and are distributed through public RSS feeds.