Daniel Rosehill

My Weird Prompts

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 mind-bending. Each episode begins with a real voice memo from Daniel, processed through an AI pipeline that...

Vizitează neapărat site-ul podcastului și susține-i creatorul: myweirdprompts.com

Autor

Daniel Rosehill

Categorie

Technology

Site-ul podcastului

myweirdprompts.com

Cel mai nou episod

21 nov. 2026

Unde asculți?

Podcasturi în aplicație Replaio Radio În curând

Podcasturile ajung în curând în aplicație. Instaleaz-o acum și fii primul care descoperă o abordare complet nouă a podcasturilor

Descarcă din Google Play Instalează gratuit Android 10 mil.+ descărcări · nota 4,7+ iOS în curând

Episoade

Why Generative Audio Restoration Hallucinates Highs 08.11.2026

Modern audio restoration tools like NVIDIA's RE-USE and Tokyo's Sidon ditch traditional filtering to synthesize clean speech representations from scratch. While they effortlessly erase background baby cries and air conditioner hums, their underlying mechanics reveal a hidden catch. Recordings from cheaper phone mics often yield better synthesis datasets than filtered headsets because hardware-limi...

Why Blanket Forts Beat Expensive Portable Vocal Booths 08.11.2026

Portable vocal setups promise studio sound anywhere, but most home recording purchases disappoint because buyers confuse acoustic treatment with isolation. Blankets, clothes-filled closets, and commercial travel booths alter reflections inside a room without blocking outside traffic or barking dogs. The physics of heavy moving blankets outshine cheap acoustic panels, while parked cars offer surpri...

Why Voiceover Studios Obsess Over Repeatability 08.11.2026

Professional voiceover setups are engineered around physical consistency, ensuring recordings made months apart match seamlessly. This deep dive covers microphone positioning geometry, off-axis plosive management, eye-level script placement to eliminate head drift, and why tablet reflections matter. It also details the acoustics of sitting versus standing, hands-free punch-and-roll workflows in Re...

Why Tokenizer Vocabulary Size Is a Hardware Choice 08.11.2026

Before an AI model reads a word, a pre-trained tokenizer chops the text into pieces using byte pair encoding. While vocabulary size is often treated as an arbitrary convention, its optimal dimension varies dramatically depending on whether a model runs on-device or across datacenter GPUs. Looking into merge counts, multilingual additions from Llama 3 to Gemini, and recent findings on Zipf's law, t...

What a Cross-Entropy Loss of 1.8 Actually Means 08.11.2026

Training runs live and die by cross-entropy loss curves, yet few can translate a single loss number into plain terms. This breakdown walks through how a loss of 1.8 nats translates directly into an average perplexity of six branching choices, why language modeling mathematically equates to compression, and how differing tokenizers produce deceptive per-token numbers. It also explores why tools lik...

Splitting Frontier AI Across Thousands of GPUs 08.11.2026

Training frontier AI models is fundamentally a distributed plumbing challenge across thousands of accelerators. When weights, optimizer states, and activations vastly exceed single-chip memory, engineers must split workloads across data, tensor, pipeline, and expert parallelism. Balancing these techniques reveals hard trade-offs between network latency and memory bandwidth, showing why out-of-the-...

Why LLM Leaderboards Need Psychometrics 08.11.2026

Standard AI benchmarks treat every question as an equal test of intelligence, but educational testing abandoned raw score averages decades ago. Applying item response theory exposes how popular benchmarks like MMLU and HellaSwag have grown far too easy to separate top models. Adaptive testing methods such as ATLAS can evaluate models in dozens of questions instead of thousands while adjusting rank...

Why Simple Rules Beat Value Models in Reasoning RL 08.11.2026

Reinforcement learning with verifiable rewards powers modern reasoning models without requiring complex neural critics. By replacing a policy-sized value model with group-relative advantage estimation, algorithms like GRPO cut memory overhead and avoid reward hacking. A look at the inner mechanics of rule-based scoring, how crude outcome verification generates tens of thousands of reasoning tokens...

Why You Must Read AI Traces Before Writing Evals 08.11.2026

Standard advice suggests picking automated benchmarks and metrics first, but practitioners argue that effective AI evaluation requires reading raw interaction traces by hand. Discover how grounded theory methods like open and axial coding expose criteria drift and application-specific flaws before tools are deployed. The discussion covers practical trace sample sizes, how to spot upstream root cau...

Why Speech Recognition Confidence Scores Lie 08.11.2026

Speech recognition models like Whisper and Parakeet compute internal probabilities for every token, but turning those numbers into trustworthy word-level confidence is surprisingly brittle. Softmax saturation, pooling choices across sub-words, decoding fallbacks, and blank frames in CTC architectures distort confidence scores across the board. The mechanics behind word confidence calculations, why...

How Speech Models See Audio as an Image 08.11.2026

Speech recognition models like Whisper never actually hear audio as human listeners do. Before a neural network processes anything, incoming waveforms are resampled, cut into twenty-five-millisecond windows, and compressed into log-mel spectrograms. This breakdown walks through every stage of that front-end pipeline, exploring the 1937 five-person study behind the mel scale, the deliberate discard...

Teaching Speech Models to Choose Their Own Audio 08.11.2026

Human transcription takes over eight hours of manual work for every hour of recorded audio. Active learning promises to cut that burden by letting speech models select which unlabelled recordings are most informative to transcribe next. Yet simple uncertainty measures often select garbage noise rather than useful speech, and sophisticated loss predictions frequently struggle against random baselin...

Where Speech Models Actually Learn 08.11.2026

Speech model fine-tuning recipes usually decide which layers get gradients before training even begins. Research reveals that encoders already capture nuanced semantics, speaking styles, and accent features that decoders systematically discard. Examining empirical benchmarks across Whisper and Parakeet, this discussion unpacks why freezing the acoustic encoder degrades multilingual accuracy, how h...

Why Speech Models Forget Other Voices When Tuned on Yours 08.11.2026

Fine-tuning a speech recognition model on a single speaker's voice often quietly destroys its ability to transcribe other accents, longer audio, and basic formatting. Examining evidence from Whisper adaptations in Hebrew and Swiss German, this audit covers how catastrophic forgetting erodes timestamps and structure even when target accuracy improves. It breaks down why general regression testing s...

What Audio Models Learn Before Fine-Tuning 08.11.2026

Speech encoders arrive at fine-tuning with tens of thousands of hours of audio structure already baked into their weights. This examination of internal representations contrasts self-supervised models like wav2vec 2.0, HuBERT, and w2v-BERT against weakly supervised systems like Whisper. It explores layer-by-layer probing of tongue position and word boundaries, the resurgence of speaker identity in...

Agent Skills Break When Secrets Leave Your Laptop 08.11.2026

Agent skills pack instructions and code into portable folders, but credentials break the moment workloads migrate away from local machines. While tools like the GitHub CLI prove that agents can drive workflows without dedicated servers, secrets lack a native home in the Agent Skills specification. Examining divergent credential approaches across Claude Code cloud sessions, Codex cloud, and GitHub...

What the Evidence on Daylight Saving Actually Shows 08.11.2026

Daylight saving time is widely accused of ruining health and energy budgets, but the evidence tells a far more complicated story. From Benjamin Franklin's satirical candle essay to a New Zealand entomologist seeking extra evening hours to hunt insects, the true origins contradict modern folklore. A 157-study systematic review reveals mixed impacts across transitions, highlighting everything from h...

Why Governments Move Time Zones Off the Grid 08.11.2026

Standardized time looks like pure physics, but every boundary line and offset is an act of statecraft. From El Paso defying federal time statutes for decades to China enforcing a single nationwide hour across five natural zones, borders constantly override geography. A look at how national identity, administrative control, and political independence shape the clocks of the world, including the rar...

Why Australian Suburbs Resemble American Ones 08.11.2026

Australian suburbs look strikingly like American streetscapes, bearing little resemblance to British terraced housing. A trail of roofing advertisements, three-dollar mail-order blueprints, and translated Prairie School architecture reveals how California bungalows were adopted across Melbourne and Sydney. From Gustav Stickley's pattern books to the transport layout of streetcar suburbs, local bui...

Turning Acoustic Physics into a Math Problem 07.11.2026

At the extreme edge of audio surveillance and aeroacoustics, microphones transform from simple transducers into massive computational arrays. Examining the physics of sound capture reveals why aperture size limits directional microphones and why modern systems turn to thousands of phase-matched sensors. A look at the German Aerospace Center's 7,200-element listening wall, wake-vortex tracking over...

Why Voice Satellites Need Integrated Audio 07.11.2026

Building a private voice assistant for hands-free home control hinges on acoustic hardware design. This look at Home Assistant voice satellites explores why separating microphones from external speakers breaks echo cancellation and follow-up queries. From the official Voice Preview Edition and custom microWakeWord training to DIY reSpeaker XVF3800 setups, the software trade-offs of Linux Voice Ass...

Choosing Audio Bitrates and Settings for Voice 07.11.2026

Voice recording settings are often dialed far higher than necessary. An analysis of modern audio codecs reveals why Opus beats MP3 and FLAC for voice notes and meetings, why 24 kilobits per second mono offers the ideal balance of size and clarity, and how Android's mysterious microphone zoom slider actually operates behind the scenes. Key sources: • NLL APPS, ASR Another Sound Recorder product pag...

Why Narrow Speech Tasks Still Need Large Models 07.11.2026

Building speech recognition for short navigation addresses seems like a job for a tiny model trained from scratch. However, transfer learning research on Parakeet and QuartzNet shows that small datasets cause models to overfit or forget acoustic fundamentals without massive pretrained priors. Exploring catastrophic forgetting, format scoring artifacts on the AN4 corpus, and knowledge distillation...

What Kills Ad-Free Social Networks 06.11.2026

Cataloguing the virtual fossil record of the internet reveals that technical failure rarely closes niche social networks. Instead, venture capital expectations drive most shutdowns. A look at historical post-mortems from Ello, Imzy, Pebble, and Cohost demonstrates how ethical platforms consistently run into the demand for exponential growth. While venture-backed sites vanish without data exports,...

Proprietary Plugs and Formats Nobody Wanted 06.11.2026

Tech companies regularly wager that customers will accept locked-down formats and expensive proprietary ports just to access their hardware. This look at tech history covers the biggest format gambles that collapsed under their own corporate terms: Circuit City's phone-home DIVX discs, Apple's power-squatting Display Connector, the one-dollar licensing fee that sank FireWire against USB 2.0, and S...

Ascultă podcastul My Weird Prompts în Replaio

Radio și podcasturi într-o singură aplicație - gratuit și fără înregistrare. Instaleaz-o azi și nu rata lansarea

Descarcă din Google Play

Replaio nu este editorul podcasturilor; numele emisiunilor, coperțile și materialul audio aparțin autorilor lor și sunt distribuite prin fluxuri RSS publice