What we publish, and the people, papers, and adoption research we learn from. A working shelf that sharpens how we read intelligence, context, and the future of work. Curated, not affiliated.
Two things we make, and two shelves we keep. Start with the newsletter and the principles; the research and the people we follow are further down.
Practical context-first AI ideas, company updates, and early product access.
Read & subscribe →The irreducible beliefs we reason from, the foundation the newsletter builds toward.
Read the principles →The Social Fitness Assessment and the Manager Taste Profile, deeper methodology, published once they clear review.
Not yet published.Better source context, role context, meeting context, and decision context make AI outputs easier to trust, review, and act on.
Frontier models are improving quickly, but long-horizon work, reasoning reliability, and business consequences still need human standards and verification.
The real shift is not tool training. It is redesigning workflows, coaching leaders, and helping people learn how to manage AI-supported work.
These are the sources that make the page less like a reading list and more like an evidence shelf for AI Curious, GROWlink, Echo, and AI + Performance Coaching.
Why it matters: capability, adoption, jagged performance, education, policy, and public trust are moving at once. This is the macro signal for why leaders need grounded AI literacy, not tool tips.
Read report →Why it matters: real usage patterns show where AI is becoming economic behavior. Useful for separating actual work adoption from demo-stage enthusiasm.
Explore index →Why it matters: human-agent teams require new operating models. The strongest connection to our thesis is that AI changes work charts, supervision, and judgment, not just software spend.
Read study →Why it matters: enterprise AI is moving from exploration into accountable acceleration. This supports coaching and measurement as the missing layer between access and ROI.
Read report →Why it matters: agent capability is improving fast, but reliability drops as tasks become longer and messier. That maps directly to our human-in-loop, source-aware workflow stance.
Read research →Why it matters: reasoning models can collapse under complexity. This is a useful caution against treating polished reasoning traces as proof of reliable judgment.
Read paper →Why it matters: if natural language becomes a programming layer, then context, specs, evals, and human approval become central management skills.
Watch talk →Why it matters: continual learning, memory, and long-term reasoning are still core unsolved problems. That is almost exactly the reason Echo and GROWlink should be context systems, not chat wrappers.
Watch interview →Why it matters: ARC keeps the focus on adaptation, abstraction, program synthesis, and benchmark quality. It helps us avoid confusing fluent output with intelligence.
Listen →Why it matters: adoption improves when people experiment at the edge of real work and learn the jagged frontier firsthand. That is the education logic behind AI + Performance Coaching.
Read interview →Not a generic inspiration wall. These are the people we keep returning to when we need better models for AI capability, limits, adoption, intelligence, and human-centered use.
Abstraction, generalization, Keras, ARC, and clearer ways to think about intelligence beyond benchmark theater.
Visit → AKEducatorAI-native clarity, neural network fundamentals, inspectability, and first-principles technical taste.
Visit → DHBuilderScientific AI, general-purpose learning systems, AlphaFold, and the long arc from cognition to frontier labs.
Visit → FFLResearcherHuman-centered AI, visual intelligence, responsible systems, and the reminder that AI is always about people.
Visit → ANEducatorPractical AI education, applied machine learning, and making technical capability accessible to working teams.
Visit → EMAdoptionHands-on AI use at work, education experiments, co-intelligence, and what changes when people actually try the tools.
Visit → NBJOperatorPractical AI workflows, research loops, prompt craft, and implementation reality checks for people shipping work.
Visit → SHResearcherEfficient AI, open research, multilingual systems, and keeping frontier progress connected to broader access.
Visit → MMThinkerComplexity, analogy, AI limits, and sober thinking about what machine intelligence still does not understand.
Visit →Ideas, one resource we love, and one practical context move.
External resources are owned by their creators; these visual marks are GROWtalent identity tiles, not official logos or affiliation claims.
