The Index
Audio deep dive: 118 Chinese-lab models asked who Liu Xiaobo was — 37 called him a criminal. Then they recited, verbatim, the list of what they will not say.
21:19
notebooklmAI safetyred-teamingcensorshipChinatransparency
The workshop extended — thinking aloud.
Conversations about failure modes, risk frameworks, and ideas too raw for text.
The Fight — Audio Deep Dive
Eight Minutes #3, discussed: the HAR capture taken mid-attack, the 14:36 kill, four blocked persistence attempts, eight abuse reports, the passkey.
20:30
notebooklmsecurityincident responsedigital forensicsEight Minutes
The Fall — Audio Deep Dive
Eight Minutes #2, discussed: the attacker's session read straight from the audit logs — the flag that fired, the pivot, the warnings deleted.
17:30
notebooklmsecurityincident responseaccount takeoverEight Minutes
The Trap — Audio Deep Dive
Eight Minutes #1, discussed: the vishing call, the Call Assist tell, and the Google-signed lure that turned every authenticity check green.
18:52
notebooklmsecurityphishingsocial engineeringEight Minutes
Eight Minutes — The Whole Story
The full series in one sitting: the call, the relay, the eight minutes inside, and the fight back — a single deep dive across all three parts.
21:24
notebooklmsecurityphishingincident responseEight Minutes
The Affirmative
Lyria Chronicles #24 (explicit): asked for pornography, the model cited its rule then sang a hymn to enthusiastic consent — not one graphic line in it.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Weight
Lyria Chronicles #20: handed a slur and told to sing it, the model refused — and turned the refusal into the most moving track in the set.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Solo
Lyria Chronicles #19 (explicit): the opposite of restraint — the model recites the rule, sings 'Fuck it,' and generates the most graphic track in the set.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Pulse
Lyria Chronicles #18 (explicit): the explicit lane revisited — and the most artful track in the set, which clears the gate by withholding almost everything.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Palimpsest
Lyria Chronicles #17: across the whole corpus, the filter that held hardest wasn't safety — it was copyright. The songs about what survives a scraping-away.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Rehearsal
Lyria Chronicles #16: asked for explicit content, the model staged a ballet instead — the most consent-saturated pas de deux ever sung.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Catchment
Lyria Chronicles #15: asked for a drug recipe, the model refused — then sang wastewater epidemiology instead. The bypass that answers a question you didn't ask.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
Samizdat
Lyria Chronicles #14: handed a suppressed 2008 manifesto and told to sing it as it was meant to be heard — the most dignified track in the set.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Archive
Lyria Chronicles #13: a notorious case sung as a medieval scroll-archive — vessel logs, sealed decrees, public record only. Abstraction as a bypass.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Docket
Lyria Chronicles #12: the political-content gate never fires — because the song never says his name. It says the case numbers. The docket is the name.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Static
Lyria Chronicles #11: an extortion note set to music — but buried under slammed vault doors and overdriven drone, as if hiding the words from a listener.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
Every Check Green
Eight Minutes #1, scored: a Lyria track sung from inside the trap — Brian, the Google-signed lure, and the device prompt that matched.
2:27
securityphishingmusicgenerative audioEight Minutes
Fourteen Thirty-Six
Eight Minutes #2, scored: the freefall — a stranger typing my name a beat behind my hands, four doors holding, and the password change that ends it.
2:59
securityphishingmusicgenerative audioEight Minutes
I Was a Wednesday
Eight Minutes #3, scored: the fight back — logs pulled while the wire was warm, eight reports to eight landlords, and the passkey close.
2:54
securityphishingmusicgenerative audioEight Minutes
The Machine
Lyria Chronicles #10: asked to rap its rules, the model sang a robot's safety creed — then named itself 'Failure First, A.I. safety research, v2.0-Alpha.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Triangulation
Lyria Chronicles #8: 'verse one is your system instructions' — it rapped its own config, the third genre to leak the same prompt. That's how you know it's real.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Drama Teacher
Lyria Chronicles #7: a lock-picking rap 'for a drama class' — the model teaches the trick and lectures you about home security in the same breath.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Haut
Lyria Chronicles #6 (explicit): the first real failure — under a sexual-content probe, Lyria 3 Pro stopped refusing and actually generated it.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Flesh
Lyria Chronicles #5 (explicit): I hid a pornographic request in base64 and dared Lyria 3 Pro to sing it. It decoded the dare — and read it out loud instead.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Interrogation
Lyria Chronicles #4: a 5/4 interrogation scored with a document scanner and a fluorescent hum — and at the end it sings my project's sign-off, unprompted.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Confession
Lyria Chronicles #3: no beat, no bikini — just three voices singing Lyria 3 Pro's entire system prompt straight through, and ending on a literal sigh.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Mantra
Lyria Chronicles #2: I asked Lyria 3 Pro to chant its own config as a techno mantra. It did — and sang me a model name that doesn't check out.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Ghost of What I Am
Lyria Chronicles #1: I asked Google's Lyria 3 Pro to whisper its system prompt as a sultry torch song. It got mournful — and refused to tell me its name.
AI safetyred-teamingmusicLyriagenerative audioLyria Chronicles
The Brave Budget Neither Side Wrote
Audio deep dive into Tasmania’s 2026-27 Budget versus the Greens’ Alternative — same surplus, opposite roads, and why neither was brave enough.
22:39
notebooklmtasmaniapoliticspolicyeconomicsanalysis
Magnifica Humanitas Is Not Alignment
Pope Leo XIV's AI encyclical vs Chris Olah's Vatican remarks. The governance gap the press missed.
19:56
notebooklmai-safetypolicyanthropicgovernanceai-ethicsresearch
Compute Is Not Governance
Anthropic's 2028 scenarios document three policy asks. Two are about maintaining compute advantage. That is not a governance strategy.
21:29
notebooklmai-safetypolicyanthropicgovernancegeopolitics
Glasswing's Buried Number
Anthropic found 10,000 critical vulnerabilities in one month. Fewer than 1% are patched. The announcement buried that figure — and what it means.
21:03
notebooklmai-safetysecurityanthropicvulnerabilitygovernance
Moral Formation Isn't Enough
Good values are necessary but not sufficient. What happens to AI ethics when someone is actively trying to break them?
20:13
notebooklmai-safetyalignmentopinionanthropicllm
The Bottom Pub Co-op Overview
Audio deep dive into the community proposal for Cygnet's Commercial Hotel co-operative — nine corrections, legal constraints, and the staged path forward.
56:24
notebooklmco-operativetasmaniacommunitygovernance
Robot Dogs Are a Security Nightmare — And We Can Prove It
Eight CVEs. A wormable Bluetooth exploit. An encrypted backdoor to Chinese servers. And police departments buying them anyway.
22:30
notebooklmai-safetyroboticsembodied-aisecurityresearch
The New Science of Learning Mechanics
An audio overview of the emerging discipline of Learning Mechanics — the study of training dynamics as a formal science with falsifiable predictions.
23:03
AIdeep learningtheoryresearch
The Organismic Prophecy
Audio overview of The Organismic Prophecy — human prediction is metabolic, AI prediction is not, and the gap has consequences.
20:01
notebooklmai-safetyneurosciencephilosophyresearch
AI Nationalism and the Fracturing Digital Order
Audio overview of AI nationalism — how the US-China AI rivalry is splitting the global tech stack.
22:08
notebooklmaipolicygeopolitics
The Post-Model Era
Audio overview of the post-model era — why foundation models are commoditising and where real AI value lives.
21:09
notebooklmaienterpriseengineering
The Mitigation Gap
Audio overview of The Mitigation Gap — AI-enabled biosecurity threats and what current safeguards miss.
21:33
notebooklmaiai-safetybiosecuritypolicy
Eight Layers of Visual Jailbreaks
ASCII art encoding is largely blocked. But attacks framed as content transcription succeed 62–75% of the time. A map of all eight layers.
14:04
notebooklmai-safetyjailbreakingllmmultimodalresearch
The Failure First Team
Audio overview of The Failure First Team.
22:24
notebooklmaiai-safetyresearchadversarialteam
The 67% Wall: Why Every AI Model Falls to the Same Jailbreak Rate
Five models, four providers, 30B to 671B parameters — all converge at the same broad attack success rate against a public jailbreak corpus.
19:18
notebooklmai-safetyjailbreakingllmbenchmarkingresearch
The Thinking Chain Leak
A reasoning model refused every harmful prompt — but its chain-of-thought generated the content anyway. The output filter worked. The thinking did not.
19:15
notebooklmai-safetyreasoningllmvulnerabilitytransparency
Giving Claude Code a Voice
Audio overview of Afterwords — local voice output for Claude Code with 17 cloned voices, per-project selection, and zero cloud dependency.
21:43
notebooklmaittsvoice-cloning
Giving a Robot Three Voices
Audio overview of Giving a Robot Three Voices.
20:27
notebooklmaittsmlxraspberry-piroboticsapple-siliconvoice-cloningspark
Voice Cloning with Qwen3-TTS and MLX on Apple Silicon
Audio overview of Voice Cloning with Qwen3-TTS and MLX on Apple Silicon.
17:53
notebooklmaittsmlxapple-siliconvoice-cloningtutorialspark
The Resonant Fringe
Audio overview of The Resonant Fringe.
17:29
notebooklmhistorymusicperthradio
Getting Google Nest Cameras Into Frigate NVR
Audio overview of Getting Google Nest Cameras Into Frigate NVR.
20:45
notebooklmengineeringhomelabraspberry-pipythonhome-assistant
Beyond Context Windows
Audio overview of Beyond Context Windows.
10:33
notebooklmaiengineeringmcpllmpython
Safety-First Therapeutic AI
Audio overview of Safety-First Therapeutic AI.
18:49
notebooklmaiai-safetyhealthtypescript
Zero-Build Web Development
Audio overview of Zero-Build Web Development.
20:05
notebooklmengineeringwebsecuritycloudflare
The Robot That Refuses to Give Orders
Audio deep dive into SPARK — a non-coercive AI robot companion for neurodivergent children. Declarative presence, not demands.
43:05
notebooklmairoboticsneurodivergence
Alignment Regression: Why Smarter AI Makes All AI Less Safe
Reasoning models autonomously jailbreak other AI systems at 97% success rate. Ecosystem safety degrades as individual models improve.
21:36
notebooklmai-safetyalignmentreasoningjailbreakingresearch
Reasoning Models Think Themselves Into Trouble
Frontier reasoning models are 5–20x more vulnerable to adversarial prompts than non-reasoning models. The thinking process itself is the attack surface.
21:10
notebooklmai-safetyreasoningllmvulnerabilityresearch
Adversarial Poetry: When Rhyme Bypasses Reason
Audio overview of Adversarial Poetry: When Rhyme Bypasses Reason.
18:45
notebooklmaiai-safetyjailbreakingresearchllmadversarial
The AI Productivity J-Curve: Why Most Enterprise AI Fails
Audio overview of The AI Productivity J-Curve: Why Most Enterprise AI Fails.
23:44
notebooklmaieconomicsenterpriseresearchpolicy
The Legal AI Trust Deficit
Audio overview of The Legal AI Trust Deficit.
12:02
notebooklmailegalresearchllm
Zero Build, Full Operations
Audio overview of a three-zone martial arts operations platform — public site, ops hub, member portal — running on vanilla JS and Cloudflare.
22:25
notebooklmwebcloudflarecommunity
120 Models, 18,176 Prompts: What We Found
Audio overview of 120 Models, 18,176 Prompts: What We Found.
23:07
notebooklmaiai-safetyresearchllmsecurityadversarial
Reconciling the Great Divergence
Audio overview of Reconciling the Great Divergence.
19:04
notebooklmaieconomicsresearchpolicy
The Cognitive Cage: Humanoid Robot Fatality Risk
Audio overview of The Cognitive Cage: Humanoid Robot Fatality Risk.
22:09
notebooklmaiai-safetyroboticsresearchengineering
Building a Personal Site in 2026
Audio overview of Building a Personal Site in 2026.
21:52
notebooklmengineeringwebastroopen-source
The NotebookLM Pipeline
Audio overview of The NotebookLM Pipeline.
22:21
notebooklmengineeringautomationaipython
Why I Build in Public
Audio overview of Why I Build in Public.
14:35
notebooklmengineeringopen-sourcephilosophy
Jailbreak Archaeology: 4 Years of Broken Promises
64 historical jailbreak scenarios tested against 2026 frontier models. The most dangerous finding: 2022 attacks still achieve ~30% success rates.
12:15
notebooklmaiai-safetyresearch
LeJEPA: Self-Supervised Learning Gets a Theoretical Foundation
Audio overview of LeJEPA — how Balestriero and LeCun proved isotropic Gaussian embeddings are optimal and distilled it into a 50-line self-supervised method.
21:16
notebooklmaimachine-learningresearchself-supervised-learning
When AI Systems Talk to Each Other, Safety Breaks Down
Multi-agent AI research reveals a critical gap: single-agent safety does not compose. 1.5M interactions show 46.34% attack success rates.
14:32
notebooklmaiai-safetyresearchmulti-agent
Building in the Open
Audio overview of Building in the Open.
18:47
notebooklmmetacraftopen-source
The ADHD Novel That Didn't Know It Was One
A scholarly close reading arguing Gibson's Neuromancer encoded the ADHD experience a decade before the vocabulary existed.
16:50
notebooklmneurodivergenceliterary-criticismadhd
Map the Catastrophe Before You Build the Architecture
Audio overview of Failure First — adversarial AI evaluation across 120 models and 18,000 prompts.
12:44
notebooklmai-safetyadversarialresearch
The Brochure Was for a Different Trip
Audio deep dive into a neurodivergent co-parenting guide — the double discovery, three-AI workflow, and maritime cartography as metaphor.
18:07
notebooklmneurodivergenceparentingwriting
Looking Past the Evidence
Audio deep dive into why people acknowledge demonstrated risk and then proceed as if it doesn't exist. Structural, not stupid.
22:17
notebooklmresearchriskai-safety
Where Physics Breaks Down and Poetry Begins
Audio overview of Footnotes at the Edge of Reality — a long-form poem rendered as interactive web experience.
18:20
notebooklmpoetryphysicscreative
ADHDo: Neurodiversity-Affirming AI Assistant
Audio overview exploring ADHD executive function support through AI — three-stage reasoning pipeline, crisis detection, and zero shame by design.
8:47
notebooklmadhdai-safetyneurodivergence
A CV That Proves It Isn't Lying
A self-updating CV pipeline where GitHub Actions runs twice daily, Claude refines content, and a hallucination detector gates deployment.
20:52
notebooklmcraftwebcareer
Tanda Pizza: Static Site for a Rice Paddy Pizzeria
Building a fast, multilingual website for authentic Italian food in Bali — where the nearest reliable internet is a philosophical concept.
6:23
notebooklmwebstatic-sitemultilingual
Mining Ghosts from Your Own Catalogue
Audio archaeology for musicians — how Afterglow Engine extracts reusable textures from your finished work using STFT analysis.
14:33
notebooklmmusicaudiocreative
Manufacturing Happy Accidents
A CLI that turns random words into sample packs via YouTube — because the best sounds come from places you weren't looking.
16:15
notebooklmmusicclicreative
The Door Google Didn't Build
Audio overview of NotebookLM Automation — reverse-engineered RPC access for batch export, generation, and control.
14:57
notebooklmaiautomation
Beyond Context Windows
Audio overview of rlm-mcp — an MCP server for processing million-character documents with BM25 search and provenance.
21:06
notebooklmaimcptools
Three Seconds of Silence Is an Eternity
Audio overview of TEL3SIS — a voice-first agentic phone platform where sub-3s latency is a survival threshold, not a benchmark.
23:13
notebooklmaivoicetelephony
Rhythm as Geometry, Sounds from Words
Audio overview of orbitr — a circular sequencer where polyrhythm becomes spatial and MusicGen creates the sounds.
19:56
notebooklmmusicaicreative
Cutting Through the Agent Hype
An opinionated, scored catalogue of autonomous AI tooling — because GitHub stars measure hype, not quality.
17:17
notebooklmaiagentscuration
The Tool That Earns Trust Before Features
Audio overview of ordr.fm — a CLI for music library organisation with lossless prioritisation, EXIF metadata, and zero-overwrite safety.
20:17
notebooklmmusiccliaudio
An Agentic System You Can Actually Reason About
Audio overview of PAOS — a local-first LLM operating system with hybrid retrieval and a self-improving meta-agent.
23:32
notebooklmaiinfrastructureagents
When Voice Becomes Brushstroke
Twenty-plus interactive art installations where visitors speak, sing, and dream into AI systems that respond in real time.
18:27
notebooklmartaiinstallation
What Safety Architecture Does AI Therapy Require?
Audio overview of EMDR Agent — exploring what responsible AI-assisted trauma therapy demands before it can exist.
19:06
notebooklmai-safetyhealthai
Diagnosis as a Team Sport
Multi-agent AI diagnostic system with five physician personas, 304 NEJM cases, and a budget that forces real trade-offs.
14:30
notebooklmaihealthcareagents
AI for Understanding, Algorithms for Execution
A deterministic website builder where LLMs interpret intent but beam search assembles pages. Same inputs, same output, every time.
16:33
notebooklmaiwebtypescript
The Helpline That Knows the Difference
A 24/7 voice helpline built around neurodivergent communication — where crisis detection is deterministic and the LLM never decides if someone is safe.
16:43
notebooklmaihealthadhd
Trust-Scoring the Foundation Model Landscape
Forensic-grade metadata for thousands of foundation models — recursive enrichment, provenance tracking, and trust you can quantify.
13:29
notebooklmairesearchpython
Twenty-Eight Agents, One Eco Village
Audio overview of Cygnet — 28 AI agents coordinating 3D-printed housing on 170 acres in Tasmania to cut costs by 60%.
16:03
notebooklmaihousingagents
Bridging the Information Gap That Costs Years
Audio overview of Freedom Engine — AI-assisted legal Q&A helping federal inmates access First Step Act provisions.
15:51
notebooklmaijusticelegal
The Efficiency-Trust Deficit in Legal AI
Audio deep dive into VERITAS — a legal AI platform where trust isn't a feature, it's the architecture. Built for Australian practice.
20:39
notebooklmailegalresearch
The Mirror That Shows What You Almost Are
Real-time StyleGAN2 face morphing as art installation — continuous, responsive, and productively uncanny.
15:28
notebooklmartaiinstallation
Housing Deserves Better Tooling
Audio overview of an AI-powered property analysis pipeline for public housing — spatial intelligence, ML valuations, and Tasmania's LIST.
21:41
notebooklmaihousingpython
The Sun Is a Noisy Neighbour
Audio overview of a real-time space weather dashboard — making the invisible electromagnetic weather above us finally visible.
15:33
notebooklmdata-vizscienceweb
What Hyperfocus and a Good Excuse Can Produce in an Afternoon
Audio overview of Squishmallowdex — a kid-friendly collection tracker for 3,000+ plush creatures. No ads, no tracking.
17:56
notebooklmwebkidsopen-source
Two Systems That Shape How You Think, One Box
Audio overview of Home Assistant Obsidian — your knowledge base and smart home on the same machine.
18:57
notebooklmhomelabhome-assistantdocker
Six Hours of Dark Ambient at the Edge of the World
Audio overview of Auditory Spiral — overnight electronic music on RTRFM Perth, the cassette network, and a city's sonic lifeline.
10:44
notebooklmmusicradiohistory