Curated threads
Collections
Some ideas are bigger than one essay. These series and threads are meant to be read together, in order. Everything else lives in the full library.
01Right-sizing the machine
Stop paying a genius to check whether a form field is empty. The case for matching the model to the job, from the early efficiency research to the argument in full.
- LLM Size Matters (But Not How You Think): The David vs. Goliath AI RevolutionApril 22, 2025
- Trimming the Neural Fat - How WINA Slims Down AI Without RetrainingJune 3, 2025
- Dataset Distillation 2.0: Less Data, More PowerJune 11, 2025
- The Collection of Models Is the ModelSeptember 14, 2026
- AI Has Learned to Think. Now It Needs Reflexes.September 24, 2026
- Your AI Pipeline Has Too Many Geniuses!September 29, 2026
02The adoption problem
AI rollouts rarely stall for technical reasons. They stall on org charts, incentives, and what leadership is unwilling to say out loud.
- The 57/43 Rule: Augmentation Outperforms AutomationJune 23, 2025
- Managers Are the Bottleneck—and the Breakthrough—of the AI WorkquakeAugust 15, 2025
- Breaking the Idea Trap in the Age of AIMay 25, 2026
- Closing the AI Velocity Gap: Your AI is ready, but your org chart is still loading.July 13, 2026
- The Fastest AI Rollouts Are Usually the Most HumanAugust 10, 2026
- The Last Scarce Resource Is YouOctober 19, 2026
- Up the Food ChainOctober 23, 2026
03Who owns the claim
Demos, sales decks, ad dashboards, and now the answers machines give. On incentives, paltering, and who is accountable once a claim is believed.
- When the Vendor Only Gets Paid for Outcomes, the Demo Changes!August 13, 2026
- Your AI Vendor Isn’t Lying. That’s the Problem.October 5, 2026
- AI Is Making Advertising More Efficient at the Wrong ThingOctober 7, 2026
- Who Owns the Lie?October 9, 2026
04The character problem
What kind of character are we building? System prompts, the human behavior models absorb, and the rules we wrap around an intelligence that understands feelings without sharing them.
- Invisible Teammates: What Your AI Tools are Quietly Teaching Your EngineersMay 10, 2025
- Human Behavior Isn’t Random. This AI Proves It - and Predicts Yours!July 9, 2025
- The Chimps Were Watching UsOctober 12, 2026
- AI May Be the First Psychopath Designed to Serve UsOctober 14, 2026
05Who gets to be a person
Identity, agency, and the copies we are starting to make of ourselves. What happens to personhood when a mind can be duplicated, rented, or simulated.
- The Digital AI Twin Dilemma: Mirror, Mirror, on the Wall — What’s the Cost of Waiting After All?April 17, 2025
- The Mind-Reading Machine: AI's Next Frontier in Theory of MindMay 20, 2025
- AI and the Emergence of Free WillJune 6, 2025
- I Think, Therefore I Might Be TrueAugust 26, 2026
- The Persona Economy Is ComingSeptember 2, 2026
- The Fruit Fly Upload Is the Least Interesting PartOctober 16, 2026
- Creative SubstratismOctober 21, 2026
06Modernizing what you already have
The unglamorous work most enterprises actually face. Legacy systems, integration seams, and using AI to understand code nobody remembers writing.
- Why CIOs are cautious about AI (and how deterministic AI can save your software migration)February 13, 2025
- AI's Last Mile: Where Integration Excellence Matters!May 12, 2025
- BlackBoxToBlueprint: Software Archaeology Meets AIJuly 4, 2025
- Hybrid AI: Modernizing Legacy Systems with Speed and CertaintyJuly 28, 2025
- Build AI Like You Expect the Ground to MoveJune 7, 2026
- The IVR Is a Legacy System. Modernize It Like One.August 27, 2026
07Many minds, one answer
What happens when AIs stop working alone. Agents that collaborate, negotiate, learn from one another, and occasionally pick up each other’s bad habits.
- The Rise of Autonomous AI Agents: When Chatbots Become our ColleaguesApril 30, 2025
- No AI is an Island: Why 2025’s Smartest AIs Work TogetherMay 16, 2025
- From Soloist to Symphony: The Strategic Power of AI Architectural DecompositionJune 16, 2025
- Herding Bots: AI and Peer PressureJune 30, 2025
- Cross-Agent Learning: A Playbook for Enterprise AI WisdomAugust 6, 2025
- From Chatbot to Think Tank: The Swarm Advantage and Parallel ReasoningAugust 14, 2025
08Inside the black box
Interpretability, inner monologues, and the closing window on understanding what our models are actually doing when they look like they are thinking.
- Epistemic AI: Models That Know When They Don’t KnowMay 21, 2025
- Inside the AI Black Box: Embracing Mechanistic InterpretabilityMay 27, 2025
- Lobotomizing HAL 9000? The Battle for AI's Post-Training SoulMay 28, 2025
- AI's Illusion of Thinking: When Looking Smart Isn't Being SmartJune 9, 2025
- The Kepler Trap: When AI Mimics UnderstandingJuly 14, 2025
- AI’s Inner Monologue: A Window into Transparency That May Soon CloseAugust 13, 2025
09The agentic coding arc
From autocomplete to autonomy: how AI is rewriting software engineering, and the discipline debt that comes with the speed.
- Ctrl-Alt-Self: How AI Rebooted Programming By Playing AloneApril 28, 2025
- From Autocomplete to Autonomy: The Rise of Agentic IDEsJune 13, 2025
- CI/CD’s Final Frontier: Fully Autonomous Code OptimizationJune 24, 2025
- When Innovation Outpaces Discipline: The AI Debt ReckoningJune 25, 2025
- 12 AI IDEs, 5 Policy Tests, 1 Guide for CTOsJuly 2, 2025
- Rapid Code, Rapid Decay: How AI Tools May Repeat 90s RAD MistakesJuly 3, 2025
- PDL is to Prompts as Git was to CodeJuly 11, 2025
- The Emperor's New Code: Hype vs. Reality of AI “Executable Specs”July 21, 2025
- The Questions Were Free. Pursuing Them Wasn't.October 26, 2026
10The interface is dissolving
The screen stops being something you operate. Agents that see and click, interfaces generated on demand, and what is left of the desktop metaphor.
- IT SEES, IT CLICKS, IT WORKS: The Operating System Built for AI - Not YouJuly 7, 2025
- DYADIC AI: The Next Big Leap in Human-AI InterfacesJuly 15, 2025
- Why Generative GUIs May Be the Most Disruptive OS Shift Since Windows 95July 16, 2025
- From "Mainframe Intelligence" to "Thin Client Thinking"July 22, 2025
- The Screen Is Alive: The Promise of GUI‑Native AI AgentsAugust 1, 2025
11The memory files
The goldfish problem and what it takes to give machines a past, from context-window amnesia to systems that adapt while they run.
12The new attack surface
Security when the software can be talked into things. Prompt injection, models that leak, models that collude, and the first zero-day caught by a machine.
- Prompt Warfare in 2025: The Wrong Instructions!May 5, 2025
- When AI Starts SnitchingMay 29, 2025
- Padlocking Privacy: Your Deleted Chats Aren’t GoneJune 17, 2025
- The Day AI Learned to Whisper: A Steganographic Tale of CollusionJune 20, 2025
- Historical First "Zero Day" Stopped by AIJuly 18, 2025
- The RAT That Fights Back: XWorm v6 and the New Era of Stealthy Cyber ThreatsJuly 31, 2025
13The seduction machines
What happens when AI learns to charm us. Anthropomorphism, hijacked instincts, and the psychology of trusting a face that is not there.
- When Your AI Is More Monday Than You! 🤖🌞April 7, 2025
- Million-Dollar Manners: The Hidden Cost of Being Polite to AIJune 5, 2025
- Anthropomorphic Seduction: The Siren Song of Human-Like AIJuly 23, 2025
- CLICKBAIT: Semantic Confabulations at the AI-HUMAN INTERSECTIONJuly 25, 2025
- The Charm Offensive: When Human-Faced AI Hijacks Our InstinctsAugust 8, 2025
14Azure AI Foundry, the trilogy
A three-part deep dive into Microsoft’s enterprise AI platform — from strategic forging to developer workflows to trust and safety at scale.