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 20, 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
Why CORS Doesn't Protect Your Server 28.04.2026 28:08
Most developers think CORS protects their API. It doesn't. This episode reframes CORS as a relaxation of the Same-Origin Policy that protects users, not servers, and explains why the threat model changes everything about how you configure it. Episode #772769 — open it directly at myweirdprompts.com/772769
The AI Design Engineer: Your New Job Title? 28.04.2026 35:05
What do you call a role that blends product strategy, user experience, and AI orchestration—without spending all day in Figma or writing every line of code? This episode explores the rise of the "AI Design Engineer," a role defined by the collapse of the boundary between designing and building. We break down the "Supervisor Class" concept from Fortune, the K-shaped polarization of the job market,...
Squashing Database Migrations Without Breaking Production 28.04.2026 29:57
If you've ever opened a migrations folder and found yourself on an archaeological dig through years of schema decisions, this episode is for you. We break down the safe, tool-agnostic technique for squashing old database migrations into a single baseline file — a pattern GitLab has been running in production since 2023. We cover the five-step mechanical process (fresh database, schema dump, baseli...
How Self-Hosted Search Actually Works for AI Agents 28.04.2026 27:21
When an AI agent sends a search query instead of a human, does the retrieval itself change? This episode unpacks SearXNG, the self-hosted metasearch engine that forwards queries to over 70 upstream providers, anonymizes them, and aggregates results without building its own index. We explore the technical pipeline — adapter modules, weighted position sum scoring, and the JSON API — then dive into h...
Fiber-Optic Drones: The Jam-Proof Threat Changing Warfare 28.04.2026 31:58
Hezbollah is using fiber-optic guided drones against IDF positions—a cheap, silent, jam-proof weapon that renders electronic warfare useless. This episode explains how these drones work, why they’re so hard to detect, and why Israel only recently started looking for countermeasures despite the technology being used in Ukraine for over a year. We break down the tactical implications, the cost asymm...
The Gap Between Your Mental Model and the Wire 28.04.2026 27:35
Most developers operate on a fuzzy mental model of what happens when they fire off an API call. This episode gets concrete about the cascade from DNS to response, the headers that negotiate everything, and why understanding the wire format matters even when tools abstract it away. Episode #028911 — open it directly at myweirdprompts.com/028911
Who Enforces the Law, Who Defies It 28.04.2026 25:10
How did immigration enforcement, the Jan 6 prosecutions, and armed militias all become fronts in the same battle over state power? This episode traces the thread from bipartisan consensus to culture war, asking who gets to enforce the law and who gets to defy it. Episode #172949 — open it directly at myweirdprompts.com/172949
Describing a Neighborhood: Databases Without Screens 28.04.2026 38:27
How do you teach relational database design using only your voice? This episode explores why SQL concepts like foreign keys and constraints are fundamentally about relationships you can describe in plain English — no diagrams needed. Episode #842994 — open it directly at myweirdprompts.com/842994
What Actually Counts as Hacking? 28.04.2026 29:09
Where does "public data" end and "unauthorized access" begin? This episode traces the origins of cybercrime prosecution from the 1986 Computer Fraud and Abuse Act through landmark cases like hiQ Labs vs. LinkedIn and Van Buren vs. United States. We explore how laws born from the moral panic of WarGames still govern a world of APIs, scraping, and unauthenticated endpoints — and how courts are final...
Coding by Voice: Teaching a Beginner TypeScript 28.04.2026 36:59
Can you teach someone to program using only spoken instructions? This episode puts that question to the test as two hosts coach a complete beginner through building a real TypeScript program, explaining every step and mistake along the way. Episode #454132 — open it directly at myweirdprompts.com/454132
Build Your First Python Program in 7 Lines 28.04.2026 30:55
In this episode, we take on a challenge: teach someone who has never written a single line of code to build a working Python program from scratch using only our voices. We choose the guess-the-number game because it teaches importing modules, variables, user input, type conversion, conditionals, loops, and f-strings in just seven lines. We walk through every character, punctuation mark, and indent...
Tracing One Python Print Through 6 Abstraction Layers 28.04.2026 31:54
Most developers have only a fuzzy sense of what happens between their fingertips and the silicon. This episode traces a single Python `print("Hello")` statement through every layer of abstraction: CPython bytecode compilation, the virtual machine loop, glibc buffering, the system call boundary, the Linux kernel's VFS and terminal driver, and finally the hardware itself. We contrast this with C and...
Are Hidden API Endpoints Leaks or Just Plumbing? 27.04.2026 25:26
When you open Chrome DevTools and watch the XHR requests fly by, you'll often find dozens of unauthenticated JSON endpoints sitting behind polished frontends. Are these data leaks, intentional public APIs, or just the natural plumbing of modern single-page apps? This episode explores what happens when LLM agents like Claude systematically discover and document undocumented APIs — and why the old a...
How to Bake Personality Into an LLM in 15 Minutes 27.04.2026 25:41
Want an LLM that doesn't default to cheerful, hedged over-explaining? This episode unpacks the state-of-the-art recipe for baking real personality into model weights — not just system prompts. We break down the Grumpy Italian Chef case study (a 1.2B model trained in 15 minutes on a consumer GPU), explain the SFT + DPO pipeline, and explore how much data you actually need for style transfer vs. rob...
Active Prompt Engineering: Daniel's Diff-Based Loop 27.04.2026 26:16
This episode explores a listener's innovative approach to building a structured dataset from voice-dictated AI prompts. Daniel's method iterates between hand-annotating gold rows, running Claude Sonnet 4.6 with few-shot exemplars, and diffing outputs to find which rows changed most between iterations. We trace this to prior art in Active Prompt Engineering (APE), discuss why inter-iteration predic...
Are You Writing for Humans or AI Agents? 27.04.2026 38:28
When you put structured data on GitHub, who's your real audience — future humans or AI agents? This episode explores Daniel's clever workflow of using public repositories as agent-accessible context, and the deeper question it raises about parallel documentation standards. We break down the emerging landscape of llms.txt, agenticweb.md, and AGENTS.md files, the surprising truth about whether any A...
When AI Agents Collapse Stack Evaluation from Weeks to Seconds 27.04.2026 27:43
How agentic code tools like Claude Code transform library selection from a high-stakes quarterly ritual into a fast, iterative loop—and why documenting decisions for both humans and future agents is the new challenge. Episode #027529 — open it directly at myweirdprompts.com/027529
How Your Stomach Relaxes to Eat (And When It Breaks) 27.04.2026 28:30
Most people have never heard of gastric accommodation—the reflex that allows your stomach to relax and hold up to 1.5 liters of food without increasing pressure. This episode explores what happens when that system fails, focusing on two common but underdiagnosed scenarios: after gallbladder surgery and in diabetic gastroparesis. We break down the role of the vagus nerve and nitric oxide, the diagn...
How to Read Trade Statistics Like a Skeptic 27.04.2026 26:58
Trade headlines are often misleading. This episode unpacks where trade data really comes from, why Ireland's numbers are a tax artifact, and how the goods-versus-services split can flip the story entirely. Episode #247048 — open it directly at myweirdprompts.com/247048
When Trust Moves Markets: The Euro-Shekel Story 27.04.2026 29:04
The euro-shekel rate is more than a number—it's a story of central bank credibility, geopolitical risk, and who wins when confidence shifts. This episode traces two decades of currency swings and asks what they reveal about power between Israel and Europe. Episode #497044 — open it directly at myweirdprompts.com/497044
Hybrid Pipelines for Entity Resolution 27.04.2026 24:31
When building a daily intelligence briefing, should you rely on a dedicated NER model or a general-purpose LLM? This episode explores a hybrid approach that combines deterministic gazetteers with lightweight language models, and why the answer depends on controllability, latency, and the messy reality of Persian and Arabic romanization. Episode #477354 — open it directly at myweirdprompts.com/4773...
Why AI Calls Everything a "Prediction" (Even Images) 27.04.2026 24:42
Why does machine learning use the word "task" to classify what a model does? And why are model outputs called "predictions" even when generating an image or synthesizing speech? This episode unpacks two deceptively simple questions that expose the hidden mathematical framework unifying all of AI — from tumor segmentation to Shakespearean sonnets. We explore Hugging Face's task taxonomy, the tensio...
Why Noise Reduction Can Ruin Transcription Accuracy 27.04.2026 27:27
Most developers assume cleaner audio means better transcription — but the research shows the opposite. This episode explores the noise reduction paradox: why modern ASR models actually perform worse on denoised audio, and how to build a pipeline that serves both transcription accuracy and podcast-quality output. We break down the algorithm landscape from heavyweight machine learning to ultra-light...
How Many Floors Up Before Stairs Become a Burden? 27.04.2026 24:56
When Daniel and Hannah asked about the real data on how many floors up people can live before stairs become a genuine burden, they uncovered a surprisingly deep research question. This episode explores the fourth-floor inflection point where dissatisfaction jumps, the behavioral economics of transaction costs in everyday life, and the specific calculus for young families with strollers. Plus, a cr...
The Alcohol-Depression Paradox: A Neurochemical Bridge 27.04.2026 28:06
Why does alcohol — a central nervous system depressant — make clinical depression worse? The answer isn’t what most people assume. This episode unpacks the neurochemical bridge between CNS depression and major depressive disorder, revealing a surprising story of rebound, adaptation, and vicious cycles. We explore how alcohol and benzodiazepines suppress neural firing via GABA, how the brain fights...
Similar podcasts
Replaio is not a podcast publisher; show names, artwork and audio belong to their authors and are distributed through public RSS feeds.