#6016: What 70 Billion Parameters Actually Means

A single artificial neuron, built from scratch — then the hard part: does that neuron analogy still describe a real transformer?

transformer-architecturefine-tuningai-inference

#6015: How Banks Move a Million vs a Billion Dollars

A million and a billion are the same number to a computer. The difference is liquidity, netting, and where the payment actually settles.

banking-and-paymentscurrency-and-exchange-ratescyberattacks

#6014: Why Every AI Model Is Secretly a Decoder

Encoder, decoder, or both? The three transformer shapes, why one won, and why BERT still refuses to die.

transformer-architecturetraining-dataai-inference

#6013: What Happens Inside One LLM Forward Pass

One trip through the network, layer by layer — and why a classifier needs one pass while an LLM needs hundreds.

ai-inferencetransformer-architecturegpus

#6012: Inside the Decode Loop: How LLMs Generate One Token at a Time

The model doesn't choose tokens or decide when to stop — the engine does. A step-by-step tour of one decode iteration.

ai-inferencegpustransformer-architecture

#6011: Inside LLM Inference: Prefill, Decode, and the KV Cache

One 8K-token sequence on a 13B model eats 6.7GB of KV cache—while the weights are only 26GB. Here's what actually happens per request.

ai-inferencetransformer-architecturecontext-and-memory

#6010: Ten Ridiculous Wearables, Ranked by Sincerity

Brainwave cat ears, a tomato-dispensing shoulder robot, and a $3,000 sock tracker — all built with total sincerity.

wi-fielectronicsproduct-durability

#6009: Vertex AI Is Now the Gemini Agent Platform

Vertex AI just became the Gemini Enterprise Agent Platform. Here's the family tree, the five products, and when to actually move.

ai-inferencelocal-aicloud-computing

#6008: Why Codex Writes Code for a Dead API

Models don't hallucinate your API — they quote real docs from the wrong year. Here's why, and what actually fixes it.

hallucinationsai-coding-assistantsmodel-context-protocol

#6007: What a $1M AI Inference Contract Actually Contains

The sticker price isn't the deal. Inside the commit floors, discount curves, and overage traps of enterprise AI contracts.

ai-inferenceprocurementcurrency-and-exchange-rates

#6006: Data Residency vs. Inference Residency: What the Law Actually Says

GDPR doesn't require local GPUs. Here's what actually drives geographically restricted AI inference — and the three layers everyone conflates.

ai-securityai-inferencegpus

#6005: What's Actually Behind a Government AI Chatbot

Singapore runs open-weight models on commercial cloud. The UK built its own chatbot, then shut it down for Copilot. What's really behind civil-serv...

ai-inferenceopen-sourcelocal-ai

#6004: Nobody Can Buy Unlimited AI Inference

Every published tier, every reserved deal, every prepaid reservation — all bounded. Even Azure says it out loud: quota isn't capacity.

ai-inferencegpuscost-of-living

#6003: What Enterprises Actually Pay for AI Inference

Twelve and a half billion dollars in enterprise API spend — and nobody can tell you who spent it. Inside AI's most undocumented market.

ai-inferencefine-tuningcloud-computing

#6002: Claude on AWS: Three Doors, One Rate Card

Same per-token price everywhere — so why buy Claude through AWS instead of Anthropic? The answer is commitments, compliance, and silicon.

ai-inferenceai-coding-assistantsprocurement

#6001: What Do Inference Providers Actually Sell?

Same open weights, 20x cost gap. What Fireworks, Together, and Groq actually sell — and when to just rent a GPU.

ai-inferencegpuslocal-ai

#6000: 6,000 Episodes and a GPU Bill: Inside My Weird Prompts

Six thousand episodes, one human in the loop, and a Modal bill that keeps growing. A behind-the-scenes update on the podcast.

ai-coding-assistantsworkflow-automationspeech-recognition

#5999: Who Teaches Actors Those Accents?

Dialect coaches break accents into six moves, not forty — and the science says imitation ability is a bell curve, not a gift.

cognitive-biaseseducationtranslation

#5998: The Worst Cars Hiding in Rental Fleets

The Kia Rio tops federal fatality tables at 170 driver deaths per million. Here's what the data says about the cars waiting in the rental lot.

cars-and-drivingconsumer-protectiontourism

#5997: Counting Our Listeners, Then Planning a Cave

What Podtrac actually measures, why downloads jumped 79% while listeners stayed flat, and how long you'd need in a cave to hear it all.

workflow-automationpodcastingdata-storage

#5994: Hebrew in Latin Letters: A TTS Pipeline Fix

A token classifier that finds Hebrew words written in Latin letters so TTS stops mangling them — built for under $3.

speech-recognitiontext-to-speechtraining-data

#5995: NVIDIA ASR: Parakeet, Canary, and Nemotron Explained

Whisper gives you one dial. NVIDIA gives you five. A practical map of the Parakeet, Canary, and Nemotron speech models.

speech-recognitionlocal-aiopen-source

#5993: Why Every Renovation Show Looks the Same

Shiplap, white quartz, open plan — why does every renovation show produce the identical room, and who's really choosing it?

constructionethicsarchaeology

#5992: Ridiculous But Legal Ways to Get RAM for AI

Memory prices are up 485% on some kits. Here are ten ridiculous — but completely legal — ways to get RAM for AI without the markup.

gpuslocal-aipc-hardware

#5991: Batch Inference on Your Own GPU: vLLM, SGLang, Ollama, llama.cpp

Hosted batch APIs sell you half-price tokens. Here's how to get the same file-in, file-out workflow on your own Linux box.

ai-inferencelocal-aigpus

#5989: Twenty Tiny Beats That Make a Podcast Stick

Twenty novel recurring segment pitches — and the test each one has to pass: does it add context, or just noise?

ai-alignmentai-evaluationprompt-engineering

#5988: Can You Engineer a Toddler's Accent?

A 15-month-old, a 20-word posh vocabulary list, and the science of whether you can actually install an accent on purpose.

language-learningchild-safetycultural-differences

#5987: The LEGO Cam: When LEGO Gave Kids a Movie Studio

It won Toy of the Year and flopped anyway. The LEGO Studios set that handed kids a camera, an editing suite, and a dinosaur.

history-of-technologypodcastingethics

#5986: Electronics Kits for Toddlers? What Actually Works

No kit exists for a 15-month-old — but the age question splits into three windows: kits at 5, breadboards at 10, microcontrollers at 13.

electronicschild-safetyeducation

#5985: Meccano at 125: What Kids Actually Learn From Bolts

Meccano just relaunched for its 125th year — but does childhood tinkering actually build lasting engineering intuition?

cognitive-biasesmemorymaterials-science

#5984: Reading a Mystery Adapter's Voltage Without a Label

An unpowered DC jack is electrically silent — so how do you find the right voltage? The answer isn't in the pins, it's in the current curve.

electronicsbatteries-and-powerhand-tools

#5983: Inside the Barcelona Breathing Cure for APD

Four versions of one treatment for abdominophrenic dyssynergia — and almost nobody outside Barcelona offers it.

medieval-world

#5982: Teaching a Voice Engine to Say "Yalla" Right

A small token classifier that spots Hebrew words hiding in English sentences — and why finding the word is the easy half.

speech-recognitiontext-to-speechai-evaluation

#5981: Batch Inference: Four Doors, Half Price

Same tokens, half the price — and almost nobody using it properly. A tour of batch inference from OpenAI to OpenRouter.

ai-inferenceopen-sourcegpus