ChatGPT
artificial intelligence chatbot developed by OpenAI
Episodes
-
#5771: MCP Connectors: Talk to Your Data vs. Open DataSame protocol, two patterns: pointing agents at your own database, and pointing them at public data — with very different results.Main topic -
#5622: Why Talking to AI Feels Easier Than Using ItChatGPT split into Chat and Work. A broken label printer shows why conversation beats menus — and what's hiding under the hood.Main topic -
#5606: Custom GPTs Retire: Skills, Plugins, and MCP ExplainedCustom GPTs retire December 11. Here's how Skills, Plugins, and MCP servers actually replace them — and what doesn't migrate.Main topic -
#5507: Inside ChatGPT's Hidden Pipeline: 8 Models Per ReplyOne chat turn isn't one model call. It's moderation, PII filtering, a tone-based router, and activation classifiers — here's the documented graph.Main topic -
#5481: Small Models, Big Guardrails: PII Detection in ChatGPTA tiny 600M parameter model is quietly scanning your ChatGPT tool calls for PII. Here's how that class of guardian model actually works.Main topic -
#5480: Writing a Personal Instruction for ChatGPTDaniel's personal ChatGPT prompt gets a line-by-line critique — and the case for shorter, sharper custom instructions.Main topic -
#5478: When AI Hands Out Your Phone NumberA chatbot gave out a stranger's real phone number. What happens when the model can't forget it?Main topic -
#5463: What Happens When You Click "Connect" on an AI PluginYou click accept, the tool runs, and a credential you never saw is doing the work. Here's who actually owns it.Main topic -
#5455: Custom GPTs Are Sunsetting: What Happens to Your ActionsOpenAI is retiring Custom GPTs — and custom Actions don't survive the migration. What that means for your workflow.Main topic -
#5426: AI Wrote His Yom Kippur PrayerA man skipped synagogue, opened ChatGPT instead, and got a prayer that knew his life. Is AI-generated prayer real prayer?Main topic -
#5775: What "Leaked System Prompts" Actually MeanAnthropic publishes Claude's system prompt — but not its accessory models. What does "leaking" a prompt actually mean? -
#5768: Who Actually Draws the Follow-Up Chips?Those suggestion chips aren't the model talking — they're a second, cheaper inference call plus a rules-based classifier. -
#5712: Can NASA Actually Build a Moon Base?NASA's Moon Base plan is a to-do list, not a physics wall — but the transport chain is the thing standing in the way. -
#5532: What Your Browser Isn't Telling YouDevTools quietly omits what privacy researchers need most. Here's how to actually tap browser traffic — and what shows up when you do. -
#5506: AI Is Way More Than ChatbotsThe word "AI" quietly became a synonym for "chatbot." Here's what the full model ecosystem actually contains. -
#5503: Inside the Hidden Image Generation PipelineThat one-click image generator is secretly a graph of many models. We reconstruct the hidden pipeline behind Gemini and ChatGPT. -
#5437: Hunting the Tells That Vanish When NamedThere's no tool for finding an LLM's verbal tics — you have to build one. Here's how keyness analysis works. -
#5412: Editing vs. Note-Taking for AI Fine-TunesHand-editing a model's output gives three training signals at once. Writing notes gives one — and a weaker one at that. -
#5355: Why AI Emails Don't Sound Like YouChatGPT's Gmail integration can now send email for you. The hard part is making it sound like you — not a committee-written average. -
#5183: Where AI Actually Sits on the Hype CycleAI has been doing real work for decades. So why do vendors still shout it from the rooftops? -
#5154: DeepSeek's Two-Endpoint PhilosophyDeepSeek quietly routed Pro traffic to Flash — and that routing change says everything about its two-endpoint strategy. -
#5136: How to Build a City Companion AIA travel guide GPT that went viral reveals the blueprint for a voice-powered AI that knows your city—and when to speak up. -
#5811: Can an Agent Actually Place Your Order?Agentic checkout protocols are real — Shopify, Walmart, Target. But the DIY parts live on AliExpress, which supports none of them. -
#5810: Standoffs, Anchors, and the Outlet Behind the WhiteboardA missing whiteboard, a metal-stud surprise, and the thread mismatch that turns one fastener into two. -
#5799: Two Agents, One Approval Gate: Wiring a Dictation WorkflowA dictated prompt becomes an episode — if one agent resolves facts, a second tightens prose, and a human approves the dispatch. -
#5762: Why 8 PM Bedtime Feels Wrong for Night OwlsA night-owl dad asks if 8 PM bedtime is "for old people" — and the circadian science on wake maintenance zones says his instinct is right. -
#5605: From Dolium to Cup: Rome's Wine Supply ChainRome had no wine bottles — just giant buried jars, throwaway shipping pots, and a hill of 53 million smashed amphorae. -
#5469: Open Source's New Map: India, Croatia, EthiopiaOne in three new GitHub developers now comes from a country outside 2020's top ten. What that shift means for open source. -
#5458: Two Linuses and the Fun in TechnologyWhat Linus Tech Tips and Linus Torvalds each model about play, mastery, and keeping tech fun as AI eats the tedious parts. -
#5393: Fine-Tuning a Model on 100 Hand-Edited AnswersYou don't need 10,000 examples to make a model sound like you. The real number is closer to 100 — if the edits are opinionated. -
#5387: Who Decides What Your Keyboard Knows?Your keyboard's word list was probably frozen in 2014. Here's who maintains the open-source dictionaries — and why Gboard's stays fresh. -
#5283: PH vs PZ: The Screw Bit Difference Nobody NoticesTwo cross-head bits sit in every driver set. One was an accident that became a standard; the other was engineered on purpose and almost nobody can ... -
#5157: When Labored Breathing in a Baby Means GoA reassuring oximeter reading isn't the whole story. How to read the work of breathing, count infant breaths, and know when to head in.