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

33:14

The Corner Store Loophole: Alcohol, Law & Addiction

My Weird Prompts ·

2026/07/07

A listener in Jerusalem watches a man buy high-strength beer every morning, shaking so badly he can barely hand over coins. His father died of alcoholism. Now he’s asking: why can a corner store legally fuel this cycle, and how does the human body survive decades of such drinking? We trace the legal gap between bars and off-premises retailers, the economics of cheap malt liquor, and the physiology of severe alcohol use disorder—including why withdrawal can kill.

en-us

31:16

From Phone Typing to Direct Agent: Fixing Headless...

My Weird Prompts ·

2026/07/07

When Daniel's home server suffered a kernel panic, ChatGPT diagnosed the problem perfectly — missing kernel modules, GRUB errors, the whole works. The repair worked, but it took hours because every single command had to be read off a phone screen and typed into a headless machine in a closet. This episode maps the three approaches to solving that bottleneck: cloud CLI tools with API key auth (like Shell-GPT), local models via Ollama for offline work, and a hybrid strategy that uses a tiny local

en-us

31:47

Translation-Safe Writing for Global Supply Chains

My Weird Prompts ·

2026/07/07

Machine translation has gotten so fluent we've stopped being suspicious of it. But that fluency masks a dangerous problem: when AI translates technical documents like RFQs and specs, errors get camouflaged. A study found 30% meaning loss in B2B technical text — and those weren't awkward translations, they were smooth sentences with wrong specs. This episode breaks down universal rules for writing that survives MT: eliminating pronouns, capping sentences at 25 words, locking vocabulary to one term

en-us

24:33

Why Your Phone Number Link Is Costing You Customers

My Weird Prompts ·

2026/07/07

Millions of small business websites display phone numbers as plain text, forcing visitors to manually copy and dial. Each extra step loses 20-30% of potential callers. The fix is a tiny HTML attribute combining the tel URI scheme with the E.164 international numbering standard. In this episode, we break down why the leading zero gets dropped, how to format numbers for global compatibility, and why a single canonical phone number unlocks SMS, WhatsApp, and VoIP channels. No more silent failure — just

en-us

43:20

DIY ISP Quality Dashboard Tools

My Weird Prompts ·

2026/07/07

Tired of ISPs saying "looks fine on our end"? This episode covers four open-source tools you can run on your own hardware to track actual latency, jitter, and throughput over weeks and months. From the powerful TIG stack (Telegraf, InfluxDB, Grafana) to the OPNsense-native SmokePing plugin, we walk through what to install, how to configure it, and how to use long-term data to hold your provider accountable.

en-us

37:15

HS Codes: The Hidden Trap in Global Trade

My Weird Prompts ·

2026/07/07

The Harmonized System promises a universal language for global trade — but the reality is far messier. In this episode, we break down why the same product can be classified differently across borders, how a frozen french fry importer overpaid duties for years and still faced penalties, and where customs brokers actually earn their keep. If you're navigating B2B importing, understanding the General Rules of Interpretation and national code extensions is the difference between smooth clearance and a

en-us

24:35

Pallet Puzzle: Picking the Right Freight Tool

My Weird Prompts ·

2026/07/07

Freight volume consolidation sounds simple — just divide cubic volume by pallet size, right? Wrong. The moment you introduce multiple SKUs with different dimensions, you're solving an NP-hard three-dimensional bin-packing problem. And the tool you pick makes an optimization decision on your behalf, often without telling you. We break down three very different approaches: PalletStacking's greedy algorithm, LoadMaster's genetic algorithm, and FreightOptix's heuristic shortcuts — and explain why they

en-us

30:26

What a Procurement Agent Actually Does for Small...

My Weird Prompts ·

2026/07/07

When Daniel needed two pallets of industrial storage boxes delivered to Jerusalem, he found two paths on Alibaba — DAP (a customs nightmare) and DDP (a false promise). The real solution? A procurement resourcing agent: a local intermediary who verifies suppliers, negotiates MOQs, consolidates orders, and handles door-to-door delivery with a bonded broker. This episode unpacks why the agent model isn't a 3PL, a freight forwarder, or a distributor — and why the lack of a standardized name for this

en-us

22:07

DAP vs DDP: The Real Cost of Customs Risk

My Weird Prompts ·

2026/07/07

When a DAP quote comes in ten percent cheaper than DDP, are you actually saving money—or buying yourself a part-time job as an importer? This episode breaks down the real cost difference between Delivered at Place and Delivered Duty Paid. We walk through bond requirements, broker fees, exam risks, and the accounting illusion that makes DAP look cheaper than it is. If you're buying B2B under either incoterm, you need to know where the hidden meters are running.

en-us

23:33

DDP Doesn't Mean Doorstep Delivery

My Weird Prompts ·

2026/07/07

DDP (Delivered Duty Paid) sounds like a white-glove service, but it's actually freight delivery with customs included. If you give a residential address, you'll face liftgate fees, undeliverable marks, and storage charges. This episode explains how to find third-party fulfillment centers that work within Cainiao's logistics network — letting you receive pallets at a loading dock for $5-15 each, with none of the residential delivery headaches.

en-us

32:27

How to Handle Chronic Badmouthing Without Making It...

My Weird Prompts ·

2026/07/07

When someone you're close to habitually badmouths a family member you know is a good person, every response feels wrong. Stay silent and you're complicit. Defend the target and you become the enemy. Try to reason and you're accused of naivety. This episode breaks down why this communication trap is so hard to escape — rooted in psychological mechanisms like splitting and triangulation — and offers concrete strategies for protecting your integrity without triggering explosive escalation.

en-us

23:47

What a $30 Box Teaches Us About Supply Chains

My Weird Prompts ·

2026/07/07

When Daniel tried to buy 40 industrial Euroboxes, he found a $0.50 factory price on Alibaba and a $30 local price — a 6,000% markup. In this episode, we crack open that $30 box to trace every layer of the supply chain: manufacturing, ocean freight, customs, distribution, pallet-breaking, and retail. What looks like price gouging turns out to be the arithmetic of making a 40-cent object available to one person who wants just 40 of them. A fascinating deep dive into why small-quantity purchasing costs

en-us

29:20

Obfuscation as Risk Management with AI

My Weird Prompts ·

2026/07/07

When telling a difficult story could destroy your life, how do you share it safely? This episode explores Daniel's proposal for an AI-powered obfuscation agent that goes beyond simple find-and-replace. We break down the quasi-identifier problem, the difference between redaction and true obfuscation, and how attention mechanisms enable context-aware entity recognition. From therapeutic disclosures to whistleblowing, we examine the engineering challenge of building a system that weighs narrative

en-us

36:32

Change Detection AI for Construction Sites

My Weird Prompts ·

2026/07/07

Construction sites change daily — different trucks, new equipment, shifting shadows. A layperson sees noise, an expert sees process, but what can AI see that both miss? This episode explores the rapidly maturing field of visual change detection, from Siamese networks to transformer-based architectures like ChangeFormer and DINOv2. We break down how foundation models have made ground-level change detection practical for hyperlocal use cases, the tradeoffs between sensitivity and specificity, and the

en-us

36:04

What Landlords Fear: Reforming Israeli Rentals for...

My Weird Prompts ·

2026/07/07

Most Israeli rental reform debates pit tenants against landlords in a zero-sum fight. But the real problem is a system that protects nobody properly — tenants face 30-day eviction notices with no stability, while small landlords with one unit can bleed for months over a non-paying tenant. This episode explores the landlord's nightmare scenarios, from professional tenants gaming eviction law to seven-month waits for empty properties. We compare Germany's balanced approach — strong tenant protections

en-us

22:09

Why AI Can’t Read Your Gmail (Yet)

My Weird Prompts ·

2026/07/07

You ask your AI assistant to find a quote in your inbox. It says no results. But the email is right there, three threads deep. This isn’t an AI reasoning failure—it’s a crisis of integration plumbing. In this episode, we unpack why ChatGPT and Gemini’s Gmail connectors are brittle and buggy, even though the Model Context Protocol (MCP) has existed for over a year. We trace three layers of failure: rate limits that force shallow searches, thread reconstruction errors from flat message data, and

en-us

30:09

Why Your AI Can't Find That Email

My Weird Prompts ·

2026/07/07

You ask your AI to find a quote in Gmail, it comes back empty, and the real reply is three threads down. This isn't a failure of AI reasoning — it's a failure of infrastructure. We dig into why Gemini and ChatGPT can't reliably read your inbox, what MCP actually does and doesn't solve, and why the gap between a working prototype and a production-grade integration is filled with unglamorous problems like OAuth token refresh collisions and rate limit pagination.

en-us

48:31

Living in the Gap: High-Rise Satire as Social Experiment

My Weird Prompts ·

2026/07/06

When a new tenant moves into a high-rise marketed as luxury, he notices the gap between the brochure and reality. Instead of shrugging, he decides to take the marketing literally—acting as if he lives on a higher plane of existence and the street is a foreign country. This episode unpacks what happens when someone deliberately exploits architectural marketing, how vertical living reshapes social identity, and whether ironic class performance reveals uncomfortable truths about the luxury label in

en-us

31:41

The Childcare Puzzle: Who Stays and Why

My Weird Prompts ·

2026/07/06

New dad Daniel asked us: who are the people who care for infants and toddlers professionally, and what makes them different from the rest of us who hit a wall after twenty minutes? We dig into the psychology of emotional granularity—the trained ability to distinguish "frustrated" from "overstimulated" in a pre-verbal baby—and how it predicts forty percent lower burnout. We also confront the economic reality: median wages below $15/hour, 46% of workers on public benefits, and a system where parents

en-us

33:51

When Religious Women Want Segregated Campuses

My Weird Prompts ·

2026/07/06

MK Limor Son Har-Melech of Otzma Yehudit has advanced a bill that would permit gender-segregated studies in Israeli universities and colleges. The stated goal is increasing higher education participation among religious women who currently avoid mixed-gender campuses. But the opposition sees it as normalizing religious gender separation in public institutions. This episode unpacks the philosophical puzzle: what does liberal neutrality require when a community's conception of the good involves

en-us

2,233 results

Similar Podcasts

Lead With AI

Tamara Nall

LessWrong (30+ Karma)

LessWrong

פשוט AI

Benny Farber

Double Tap

Steven Scott, Shaun Preece, Steven...

The Indian Startup Show

Neil Patel