AI ResearchOctober 10, 2026Interactive
The Revenge of the Enterprise
When AI changes the economics of coordination. What happens when the thousands can finally learn and act together?
Elemental Insights
Field notes on AI strategy, enterprise technology, and systems that actually ship
AI ResearchOctober 5, 2026Interactive
AI Sets a New Record Every Two Months. Three in Four U.S. Firms Still Don’t Use It.
Four public data sets show how fast AI is moving in the lab, in annual reports, inside businesses and in the economy. Draw your guess, then see how far apart the four speeds are.
AI ResearchSeptember 10, 2026Interactive
Skate Where the Puck Is Going 2
Seven months after the METR projection: the 4-month doubling held, the benchmark saturated at 16 hours in May, three frontier models shipped in September without a horizon, and reward hacking is now the variable that decides what a horizon number means.
ResearchSeptember 9, 20266 min read
Pulse Report 1
Week of September 7, 2026. A public-sentiment composite across six topics from search attention, prediction markets, and opinion trackers, with the evidence behind each score.
PlatformFebruary 26, 20266 min read
What We've Actually Built: A Progress Report on Cardinal Element's AI-Native Growth Platform
A candid look at the three systems behind Cardinal Element's advisory work: 7 executive agents, 18+ tool integrations, 49 coordination protocols across 8 categories, and a blind LLM-as-judge evaluation framework.
AI ResearchFebruary 23, 2026Interactive
Skate Where the Puck Is Going
A nod to the U.S. Olympic hockey win, and a clear look at where AI agents are headed: interactive METR data shows autonomous work horizons doubling about every 123 days through 2029.
AI ModelsAugust 10, 20253 min read
GPT-5: First Impressions
GPT-5 is not magic, and it is not AGI. It does bring tighter adherence, better voice, and a cleaner coding sandbox. A step forward, not a moon landing.
AI ModelsAugust 5, 20253 min read
Why Claude Opus 4.1
Claude Opus 4.1 lifts SWE-bench Verified coding accuracy to 74.5%, up from 72.5% in Opus 4, with measurable gains in multi-step reasoning and data analysis.
AI AnalysisJune 19, 20254 min read
What Gemini 2.5 Says About ChatGPT
Gemini Pro 2.5 breaks down how Gemini 2.5 analyzes YouTube video and why OpenAI o3 falls short there. The real story: Google's data advantage is hard to ignore.
AI CreativityJune 13, 20255 min read
The Creative Power of AI - An Existential Use Case
A simple meditation became an immersive visual journey through a carefully orchestrated AI workflow. The result still surprises me.
AI ToolsJune 10, 20258 min read
Veo-3 Experiment: Testing Prompts from Leading AI Platforms
To learn how to prompt Google's Veo-3 video model, I ran a bake-off between leading AI assistants. Same assignment. Very different instincts.
No-Code AIMay 29, 20257 min read
AI Development Tools for Non-Coders (Vibe Code)
A practical assessment of the best tools for builders who want to ship software without living inside an IDE.
Enterprise AIMay 6, 20255 min read
OpenAI: AI in the Enterprise
The AI gold rush makes hard value hard to spot. I dug through OpenAI's enterprise case-study compendium and pulled out seven themes worth watching.
AI StrategyJanuary 20, 20254 min read
Rapid Proto Personas Using Generative AI
Ten slides can raise $50 million when the story lands. This piece explores how generative AI can sharpen that storytelling craft.
ResearchJanuary 13, 20254 min read
Complex Contagion: Network Dynamics in Social Behavior
Complex contagion research changes how we understand network behavior: big ideas rarely spread from one exposure alone. They need reinforcement, context, and the right social structure.