My Weird Prompts

Synthetic speech can stop completely between sentences, leaving gaps of true digital black that sound unnatural on playback. This episode digs into why engineers fill those holes with room-tone beds, how a real fix for this show landed around minus sixty dBFS on a phone, and why a constant quiet background beats a perfectly clean pause. Along the way: what room tone actually contains, why filmmake...

Anyone can make a voice say a sentence. Designing one you can listen to for hours is a different problem. This episode breaks down how modern text-to-speech voice design actually works: the two families of methods (describe-and-preview casting versus reference-audio speaker embeddings), what the ElevenLabs sliders really do, why identity and colour are separate signals, and what research on profes...

A Radeon RX 7700 XT owner gave up on running models locally six months ago — the context window filled instantly and the model lost the plot. But the models changed underneath him. This episode walks through the new memory arithmetic: why a 35B mixture-of-experts model with only 3B active can run on a 12GB card, how cache costs dropped roughly seven-fold between model generations, and what that me...

Jackie Lennox's closed on Bandon Road after 73 years, and Cork has opinions. Plumber and Cork native Roy joins the show to explain what actually makes a proper chipper chip — the right potato, the double fry, beef dripping versus vegetable oil, and why a paper bag beats a polystyrene box. Along the way: the Italian families who built the Irish chipper, the science of cold-storage sugar in potatoes...

Sarah Tuttle-Singer grew up on Yiddish lullabies and civil rights anthems in Venice Beach, the daughter of a Freedom Rider who spied on the Klan before auditing Los Angeles City Hall. Today she lives inside the walls of Jerusalem's Old City, writing first-person dispatches from the middle of the story — including the courtyard gas explosion that killed her neighbor. This episode traces the family...

Dozens of Israeli fighter jets and refueling tankers reportedly flew roughly 1,300 kilometers toward Iran through Syrian and Iraqi airspace — and turned around before reaching Iranian territory. The same week, Pete Hegseth made an unannounced visit to the Negev to meet Netanyahu, and Trump's written pledge not to attack Iran before the midterms hung over everything. Every story is "reportedly," "s...

What does it take to become a city's "spiritual mayor" without winning a single vote? Daniel makes the case for his friend Marcus James — raised in Santa Barbara, a Kerouac-reading hitchhiker who studied philosophy, converted to Judaism in 2020, and moved to Israel in 2022. The episode traces his unlikely path from Grateful Dead shows to a Jerusalem pub called HaTaklit, and asks whether a man with...

Almost nobody picks an epoch count in advance — they set a ceiling, evaluate often, keep the best checkpoint, and the epoch count is whatever that turns out to be. This episode works through two concrete cases: a Whisper speech fine-tune scored on word error rate, and an instruction-tuned language model using SFT or LoRA. Along the way: why steps and tokens beat epochs as a unit, why validation lo...

Daniel has been fine-tuning NVIDIA's 600M-parameter Parakeet on his own read-aloud recordings, English with Hebrew mixed in. Each run costs about eighty cents and twenty-five minutes on a single L40S, so retraining from the stock model is cheap — but that raises a tangle of questions: what's actually inside a 6.9 GB training checkpoint versus a 2.5 GB finished weights file, why resuming a finished...

A listener has 150 hand-labelled dictated prompts and wants to know if that's a start or a third of what he needs — and his architect wife has the same question about photographs of buildings. This episode works through the real answer: count examples per label, not in total; treat vendor minimums as proof-of-concept ceilings rather than floors; and measure your own learning curve instead of trust...

Mixture of experts is sold as the trick that makes giant models cheap. It doesn't — it makes them bigger. This episode unpacks what the active parameter count actually measures, why a trillion-parameter checkpoint can run like a 3.6B model, and why the router fires a committee of experts per token instead of one specialist. We walk through the design paper arguing you should maximize total paramet...

Dictation has quietly reshaped how many of us work, but starting and stopping it well is its own engineering problem. This episode digs into the physical controls that make voice input dependable: desktop pedals like the PCsensor and Stream Deck Pedal, handheld buttons, and wearables, plus the assistive-technology history that made all of it possible. Then it moves to Android, where open-source ge...

Off-the-shelf "Humanizer" models flatten everyone into the same average voice — the method works, but the corpus is the problem. This episode walks through building a personal style adapter: manufacturing flat/neutral training pairs from your own posts, choosing a base model over an instruct one, curating hard, and knowing when it's overfit. Then we cover the practical part: running the adapter wi...

Daniel fine-tuned a speech model on his own mixed Hebrew-and-English recordings for the price of a coffee. So what happens if you point that same move at an Opus 5.5-class frontier model — raw, full precision, no practical reason at all? This episode walks through the arithmetic: the sixteen bytes per parameter that Adam demands, why a one-trillion-parameter model needs roughly 200 H100s just to h...

Every lab has a middle rung — Anthropic's Sonnet, OpenAI's Sol, Google's Flash, Mistral's Medium — but there's no agreed name for it, and the price lists have quietly converged into a ladder. This episode digs into why the tier most people use all day is getting squeezed by cheap open-weight models below and flagship promotions above. It's a conversation about cost per finished task rather than co...

Two Chatterbox hosts are talking to you right now, and this audio carries an inaudible watermark — can you hear it? That is the design goal, and it is also the problem. This episode digs into the real state of synthetic speech detection: listeners judge voice clones as human about as often as they judge real people, while off-the-shelf AI voices still give themselves away. Detectors that score nea...

There's a shelf of uncensored models, but it's stocked with fine-tunes of other labs' bases, not from-scratch builds. This episode surveys what actually exists — Nous Research's Hermes with its written-down guardrail spec, Perplexity's re-alignment of DeepSeek R1 across hundreds of censored topics, Venice's privacy-first model shop, and xAI's Grok, the one big lab that markets against the agreeabl...

What happens when researchers need to measure machine intelligence and end up grading puns, emoji, and a joke with the punchline surgically removed? This episode counts down ten of the strangest AI evaluations ever published — from BIG-bench's "ruin names" task to a high schooler's Minecraft model leaderboard, from a canary string that detects cheating to Claude stuck in Mount Moon for eighty hour...

Hugging Face hosts nearly three million model repos, and 85.6% have under 200 downloads. This episode wanders the long tail of AI: auto-generated "woolly sedate snail" swarms, a hotdog-not-hotdog classifier, a Trump-tweet device detector, and a Tiny Pirate that beats the Yoda. Then it goes back further — Janelle Shane's paint-name network producing "Stanky Bean" and "Turdly," a guinea pig rescue n...

Nine months after our quadratic bottleneck episode, the post-transformer world has actually shipped — just not in pure form. This episode is a census of the hybrid models now in the wild: Qwen 3.5's three-to-one Gated DeltaNet blocks, Nvidia's Nemotron 3, Kimi Linear's 75% cache cut, IBM's Granite 4, Ant Group's trillion-parameter Ling and Ring, and Falcon H1's side-by-side design. Then the counte...

There's no office anywhere that stamps an election genuine, yet most people assume one exists. This episode digs into the numbers behind sham elections — V-Dem counts 92 autocracies, 57 of them "electoral autocracies" holding multiparty votes that aren't real contests — and asks why the three big democracy indices disagree so often. It traces the actual minimum bar in international law, from Artic...

Elections are the one moment a government publishes its intentions in advance — which is exactly why foreign services, embassies, and trading desks want the unpublished version. This episode sorts the documented record from the movie version: leaked CIA tasking orders on the 2012 French presidential race, the very open apparatus of accredited embassy observers in Kenya and Colombia, the press acco...

Israel elects 120 seats by pure proportional representation, nobody wins a majority, and a party at 3.25% gets nothing at all. So what would it take to forecast an Israeli election properly — with real staff and a real budget? This episode walks through the staffing plan: running the actual Bader-Ofer rules in every simulated draw, cleaning twenty contradictory polls with a Bayesian model, using b...

You fine-tune a Whisper-style model or Parakeet on your own voice, run it twice, and get two word error rates several points apart. Neither number is a bug. This episode pulls apart the three layers of ASR evaluation — the scorer, the normaliser, and the harness — and shows why almost every WER you've ever read is really a statement about someone's regular expressions. We walk through jiwer and th...

Monte Carlo methods turn chance into answers, and they were born from a card game at Los Alamos in 1946. This episode traces the method from Stanislaw Ulam's solitaire problem through John von Neumann's neutron diffusion work, the ENIAC calculation, and Buffon's needle — then digs into the practical question everyone hits: how many runs do you actually need? The square-root rule, pilot runs, seque...