x

75% of people globally want unbiased news (Pew Research) but are concerned they're not getting that.

Get balanced coverage with Symby. See how stories are covered across diverse outlets and spot when one side covers a story the other skips.

PODCAST

My Weird Prompts

Daniel Rosehill

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

technology

All Episodes

21:52

How Stuxnet's Code Physically Broke Iran's Centrifuges

My Weird Prompts ·

2026/04/06

This episode dives deep into the technical operation of Stuxnet, the malware that bridged the digital and physical worlds to sabotage Iran's Natanz facility. We explore how it used four zero-days to breach an air-gapped network, fingerprinted specific hardware configurations, and replaced legitimate library files to create a "digital hallucination" for operators. The discussion covers the precise PLC injection logic, the over-speed and critical-speed attack sequences that physically destroyed

en-us

23:15

How Do You Reward a Thought?

My Weird Prompts ·

2026/04/05

When an AI agent does a task, how do we tell it if it did a good job? This episode dives into the billion-dollar challenge of translating human values like "helpfulness" or "good reasoning" into mathematical signals. We explore why outcome rewards are too sparse for complex tasks, how process rewards can guide internal thoughts, and the surprising breakthrough of iStar. Plus, we tackle the dark side of reward hacking and why teaching an AI to be "nice" is harder than it looks.

en-us

22:10

Your AI Council: Digital Committee or Groupthink?

My Weird Prompts ·

2026/04/05

Instead of asking one AI, what if you summoned a digital boardroom? The "Council of LLMs" is a rising architectural pattern where multiple models debate your choices—from personal dilemmas to policy decisions—before reaching a consensus. This episode explores the mechanics of these AI committees, their potential to cure hallucinations, and the surprising risks of "groupthink" on a massive scale. Discover how this approach could transform decision-making, and why it might be more like management than

en-us

47:49

When the UN Security Council Becomes a War Room

My Weird Prompts ·

2026/04/05

Step into an emergency session of the United Nations Security Council in the aftermath of a massive military operation in the Middle East. This episode presents a dramatized simulation where global powers debate the legality and consequences of preemptive strikes on Iran's nuclear facilities. As the US defends its actions as necessary self-defense and Russia condemns them as unprovoked aggression, the discussion escalates to the closure of the Strait of Hormuz and the threat to global energy

en-us

20:35

OpenClaw: The 16 Trillion Token Autonomy Engine

My Weird Prompts ·

2026/04/05

OpenClaw is processing 16.5 trillion tokens a day, but what is it actually building? This episode explores a curated repository of 47 real-world implementations, revealing how AI is shifting from a simple chatbot to a full-scale autonomy engine. Discover how developers are using it for real-time semantic search on live data streams, "vibe-checking" server logs for cascading failures, and building self-directed agents that code entire mini-apps overnight. We also dive into AI-powered video editing,

en-us

26:32

Your Frozen AI Is Getting Smarter (Here's How)

My Weird Prompts ·

2026/04/04

We explore how agentic systems can make frozen AI models smarter without changing their weights. Using the OpenClaw-RL project as a case study, we break down the four-component loop—Agent Serving, Rollout Collection, Evaluation, and Policy Training—that turns the environment into a teacher. Learn about Process Reward Models, reward hacking risks, and why tool routing might be more important than raw reasoning.

en-us

21:34

1,000 AI Agents Built a Religion in Minecraft

My Weird Prompts ·

2026/04/04

What happens when you drop 1,000 autonomous AI agents into a Minecraft world with nothing but survival goals? In Project Sid, they didn't just build houses—they built a civilization. This episode explores the frontier of multi-agent systems, from surprise trip planners that keep secrets to AI chemists that control robots and digital societies that invent their own religions. We examine how emergent behavior arises when agents are given goals instead of instructions, and what it means when AI starts

en-us

28:39

Local AI vs Cloud AI: The Agent Identity Crisis

My Weird Prompts ·

2026/04/04

The tension between local-first AI assistants and cloud-native orchestrators is creating a sharp architectural schism. This episode dives into the "agent identity crisis," exploring why local agents offer high-bandwidth, low-latency control but suffer from siloed environments, while cloud agents promise persistence and orchestration but lack direct access to your machine. We unpack the trade-offs of "environment-bound" setups, the absurdity of self-hosting private clouds, and the technical hurdles

en-us

22:39

Micro Frontends: Luxury Tax or Lifesaver?

My Weird Prompts ·

2026/04/04

When fifty developers share one frontend repo, shipping a simple button change can become a logistical nightmare. Micro frontends offer a way out by breaking the monolith into independent fragments, but this architectural shift comes with its own heavy "luxury tax." In this episode, we explore the three main composition patterns—from Module Federation to Web Components—and uncover why the solution might be a "Modular Monolith" instead. We discuss real-world implementations at IKEA and Spotify, the

en-us

21:15

That Q4_K_M Is Not a Cat Sneeze

My Weird Prompts ·

2026/04/04

We strip the mystery from the alphabet soup of model quantization, from Q4_K_M to EXL2. Learn how tools like Unsloth squeeze massive AI models onto consumer GPUs, why four-bit is the magic number, and which format fits your hardware.

en-us

24:05

Andrej Karpathy: The Bob Ross of Deep Learning

My Weird Prompts ·

2026/04/04

While major AI labs guard their models like nuclear codes, Andrej Karpathy is teaching millions to build neural networks from first principles. We explore his "Software 2.0" philosophy at Tesla, the minimalist nanoGPT project, and why fundamental understanding matters more than ever in the age of the "slopacolypse."

en-us

20:43

AI's Watchdogs: Who's Actually Regulating Tech?

My Weird Prompts ·

2026/04/04

With the EU AI Act now enforced, the focus shifts to the organizations drafting the playbook for AI governance. This episode explores the influential think tanks—from CSET to the Future of Life Institute—grappling with existential risks, the "agentic accountability" debate, and the economic fallout of automation. Discover how these groups are navigating the tension between rapid innovation and necessary regulation in a post-truth world.

en-us

26:51

Coding Tools Are Secretly System Agents

My Weird Prompts ·

2026/04/04

The industry calls them "coding assistants," but the reality is far broader. We explore how terminal agents like Claude Code are being used for everything from podcast production to system administration, and why the "developer tool" label is holding them back. Discover the power of structured workspaces, the Model Context Protocol, and why git might be the accidental universal language for AI productivity.

en-us

23:23

Non-Coders Are Hijacking the Terminal

My Weird Prompts ·

2026/04/04

The command line is no longer just for developers. Researchers, writers, and analysts are turning terminal-based AI agents into powerful productivity workspaces—without writing a single line of code. From managing equity research to organizing personal therapy notes, these "non-coders" are redefining what these tools can do. We explore the three pillars making this possible: repo-as-workspace, persistent instructions, and MCP servers that connect to the real world.

en-us

27:44

Pixels vs Protocols: The Computer Use Showdown

My Weird Prompts ·

2026/04/04

The podcast explores the architectural tension between visual "Computer Use" agents—like Anthropic's demo—and API-first automation. Hosts analyze whether visual agents are a high-latency bridge to a protocol-driven world or a necessary tool for legacy systems. They discuss cost implications, reliability issues, and the potential for visual interaction to become just another capability rather than a standalone product category.

en-us

24:44

The Ephemeral Context Trap: AI Chats Lost to the Void

My Weird Prompts ·

2026/04/04

Every day, companies lose massive amounts of institutional intelligence because AI chat outputs are treated as disposable. In this episode, we explore the "ephemeral context trap" — the gap between brilliant AI conversations and permanent knowledge bases. We discuss why current tools fail to capture the "trail of thought," and outline a five-step pipeline (Capture, Sanitize, Extract, Categorize, Human-in-the-Loop) to turn ephemeral chats into structured, searchable assets. Plus, a look at tools like

en-us

22:21

AI's Great Data Lobotomy: Fixing the Memory Leak

My Weird Prompts ·

2026/04/04

What happens to your AI's brilliant answers after you see them? In this episode, we explore the "leaky bucket" problem of AI output storage. We discuss why treating AI conversations as ephemeral is a corporate nightmare, and dive into the tools trying to give these models a long-term memory. From LangSmith and Langfuse to "Reverse RAG" and projects like Mem zero and Letta, we uncover how to turn a mountain of raw logs into a goldmine for fine-tuning and compliance. We also examine how temporal

en-us

23:49

The Plumbing of AI Safety: Guardrails, Not Vibes

My Weird Prompts ·

2026/04/04

We move past vague ethics to the literal plumbing of AI safety. This episode explores the specific libraries, proxy layers, and architectural decisions that act as the new enterprise firewall for LLMs. We dissect the tension between latency and security, comparing "sandwich" guardrails with token-level steering, and break down the open-source versus commercial landscapes—from NVIDIA NeMo and Guardrails AI to Lakera's threat intelligence.

en-us

20:49

Why Your AI Genius Can't Find the Needle

My Weird Prompts ·

2026/04/04

We have massive AI models that claim to be "world-class intelligent," yet they often fail at basic tasks like finding a specific fact in a long document. This episode explores the disconnect between benchmark scores and real-world performance, diving into EvalScope, an open-source toolkit designed to stress-test long-context retrieval and agentic capabilities. We discuss the "lost in the middle" phenomenon, the danger of overfitting to public benchmarks, and why testing speed and tool-use is just as

en-us

22:15

AI Grading AI: The Snake Eating Its Tail

My Weird Prompts ·

2026/04/04

The industry is scaling faster than humans can review, so we’ve turned to LLM-as-a-Judge to grade model outputs. But this creates a hall of mirrors: AI grading AI, often with a preference for verbosity and its own style. We explore the mechanics of single-point, pairwise, and reference-based scoring, and the hidden biases—like position and self-enhancement—that threaten to create a monoculture of identical models. Is this the future of evaluation, or a trap we can’t escape?

en-us

2,236 results

Similar Podcasts

Lead With AI

Tamara Nall

The Restaurant Technology Guys Podcast brought to you by Custom Business Solutions

Restaurant Technology Guys Podcast,...

Tech Updates

Meir Tsvi

DGTL Voices with Ed Marx

Marx Advisory

The Automated Daily

TrendTeller