coocBoard

coocBoard · cooc.ing

An idea in progress

Create Order
Out of Chaos.

A personal board that turns commodity compute into personal advantage. It layers your context — what makes you distinct, where you’re headed, what you’re building toward — onto raw AI capability, filters the world’s noise down to what you actually need, and nudges you up the learning → doing → scaling arc toward what only you can build.

  • Not a productivity app, a bookmarking tool, or another content feed.
  • But a board that helps one person figure out what they truly want — and move.

When everyone can automate, automation stops being the differentiator.

If every engineer, data scientist, and researcher reaches for the same tools and ships the same project — yet another RAG pipeline on a resume — the output is homogenized and the honest reaction is “so what?”

What actually distinguishes a person is their agency: their point of view, their taste, the thing only they would think to build. coocBoard exists to bring that agency to life.

Three frictions that compound.

  1. i

    Homogenization by recommendation

    Feeds collapse everyone toward the same trend. The timeline turns entirely to one topic, so people feel they must work on it too — not because it fits their arc, but because the system manufactured the urge. Individual direction gets overwritten by the crowd’s.

  2. ii

    Sludge

    People often know what they want, and the tools exist — yet they still can’t act. Interests pile up, time fragments, nothing gets tracked, and friction accumulates until progress stalls. Sludge is anything that stops a person moving forward.

  3. iii

    Sloppy learning & building

    Most people don’t know what is worth building. Their choices are bounded by what’s visible and easy, not by what’s original and valuable. That visibility gap is itself a fundamental problem worth solving.

A single, clean board for what defines your direction.

It models four coordinates of you — not a personality quiz, but a self-model you can see and steer, inferred from what you actually do:

Signal
What makes you distinct — your taste, your values, the questions you keep returning to.
Direction
Where you’re going — your bets, your arc, your north.
Capability
What you can do now, and your growing edge.
Context
Your season — time, energy, constraints, the phase you’re in.

The self-model is yours — legible and steerable, and it moves with you. Most personalization freezes a snapshot of you and then filters the world through it; coocBoard keeps the model living, shows you why this, now, and lets you agree or override rather than obey. Memory is just the substrate; agency is the point.

From compute to outcome.

Raw compute and raw code are commodities now — hand someone a coding engine and they build something, but it’s the same something everyone else builds. The hard problem isn’t generation; it’s figuring out what to build that gives an advantage. Each layer converts commodity capability into personal advantage:

Compute
The substrate — available to everyone, differentiating no one.
Empathy + context
Turns a generic model into intelligence about this person: their signal, direction, capability, and context.
Signal filtering
One question against the flood: what does this person need — not everything, not the trending thing.
Nudging
Persistent, gentle pressure toward something outstanding — not toward resume filler.

Cut the noise. Nudge the reasoning. Keep the arc.

An agent that cuts through noise

A background agent scans the internet on your behalf and filters ruthlessly: given what you care about, here’s the one thing worth your attention — and why. The reason always comes with the nudge, so you stay the judge. It never demands much time, and brings things to the table — curated, contextual, in whatever format you prefer, down to short-form quick bites.

Nudging reasoning, not just content

The board keeps your thinking unstuck by prompting all three modes of reasoning. When you stall, it’s usually because one mode has run dry — the board helps you switch and keep climbing.

  • Abductive Generating hypotheses — what might explain this? what could I try?
  • Inductive Generalizing from what you’ve seen and done.
  • Deductive Working out the consequences of what you believe.

Progressive arc-building

Knowledge should compound. A roadmap optimizes for coverage, and coverage is not capability — so the board tracks learning as a living progression, not a checklist. Each thing you explore builds on the last instead of being an isolated dopamine hit — guiding without rushing, keeping the arc visible.

Surfacing what’s worth building

Of everything you could build or learn, which is essential, original, and genuinely advancing — versus the default, sloppy choice everyone else makes? The board answers the question most people never ask.

Learning → doing → scaling.

The product grows with the person, in the same order a person grows:

  1. Learning

    The entry point. The board curates, filters, and builds the progressive knowledge arc — the person finds their agency and direction.

  2. Doing

    The board becomes an engine. With direction clarified, it helps the person build — projects with a real point of view, not homogenized filler.

  3. Scaling

    The engine compounds. What the person builds becomes leverage — for their career, product innovation, engineering, whatever they aim at.

The same arc holds whether the person is learning ML, innovating on a product, or doing engineering work — the board’s constant job is to keep them moving up it.

It should belong to the person.

coocBoard lives close to the person — on their own system, or on dedicated personal compute that belongs to them. Two reasons:

It works through them. A background worker holding your signal, your direction, and your ambitions has to be yours — a personal space, not a feed someone else tunes.

It evolves into an engine. The board is the starting form. The trajectory: a board → a “cook” engine → ultimately a personal coding engine running on the individual’s own system, executing on their behalf within the direction the board has clarified.

Four rules it never breaks.

Agency over automation
You can’t automate everything, and shouldn’t want to. The board amplifies what the person uniquely contributes; it never replaces their judgment.
Clean and uncomplicated
No over-engineering, no contemplation traps. The person should feel in control — never trapped or overwhelmed.
Order out of chaos
The name is the function: take the scatter of interests, tools, feeds, and ambitions and give it one coherent surface to act from.
Anti-sludge
Every feature faces one test: does this remove friction from progress, or add it? The measure is whether you moved — never whether you scrolled.

A product, and a research instrument.

coocBoard does several jobs at once:

  • Research — personalization, memory, and identity modeling as live AI problems: how do you model a person well enough to increase their agency rather than narrow it?
  • Human–AI coexistence — coexistence defined concretely: AI helping humans progress, not just automating their tasks. The counter-force to homogenization.
  • A real product answer — individuals and organizations face the same sludge and hype-chasing. A working coocBoard is proof the problem is solvable.

Research alignment

coocBoard lands on specific open problems in Anthropic’s Institute agenda. The agenda explicitly calls to “design interfaces that improve and promote human agency” — coocBoard is one of those interfaces. It directly counters “group epistemology when populations use the same models” (homogenization) and “degradation of critical thinking” (the reasoning-nudges), and speaks to “worker agency preservation” as AI “absorbs foundational tasks.”

The agenda’s posture is to study and measure AI’s impact from inside a lab. coocBoard is the complement — a product intervention, and an instrument to test the same questions at n=1: does an agency-promoting board measurably counter homogenization? Do reasoning-nudges slow critical-thinking decay? Does it help someone rebuild differentiated expertise once the foundations are automated?

The need is visible in the data.

coocBoard’s thesis isn’t a hunch. Across top-tier research and reporting, one pattern keeps recurring: agency is the upstream variable — and jobs, wellbeing, creativity, and learning are its downstream expressions.

What people want from AI

  • What people want from AI — 81,000 interviews

    Anthropic · 2026

    The largest multilingual qualitative AI study ever — 80,508 people across 159 countries. After unreliability, the two biggest concerns are autonomy & agency (21.9%) and cognitive atrophy (16.3%), with educators reporting atrophy 2.5–3× more than average. The conclusion: people want AI to expand human agency, not replace it.

The résumé signal is collapsing

Students are losing the foundations

What the cognitive research shows

  • Your Brain on ChatGPT: accumulation of cognitive debt

    MIT Media Lab (Kosmyna et al.) · 2025 · arXiv:2506.08872

    An EEG study where students who wrote with ChatGPT showed the weakest neural connectivity, more homogeneous essays, and reduced originality — effects that persisted after the tool was taken away. The authors call it “cognitive debt.”

  • The Impact of Generative AI on Critical Thinking

    Microsoft Research · CHI 2025

    A survey of knowledge workers finds higher confidence in AI correlates with less critical thinking — the effort gets offloaded rather than exercised.

What top minds are saying

  • Automating creativity

    Ethan Mollick (Wharton) · 2023

    In idea-generation experiments, “AI ideas were generally judged to be more similar to each other than ideas generated by people” — homogenization at the source.

  • The biggest problem of the AI age is adjustment, not unemployment

    Tyler Cowen, via Fortune · May 2026

    The economist’s read on who wins: “the people who are best at taking initiative, figuring out how AI works… doing something new.”

The view from the top

  • Machines of Loving Grace

    Dario Amodei · Oct 2024

    Anthropic’s CEO argues powerful AI should compress progress in health, the economy, and human flourishing — if it is pointed at the right ends.

  • Behind the curtain: A white-collar bloodbath

    Axios · May 2025

    Amodei warns AI could eliminate up to half of entry-level white-collar jobs within one to five years — making “what is still worth being good at” the defining question.

The economics of augmentation

  • The Anthropic Economic Index

    Anthropic

    AI use already skews toward augmentation (57%) over automation (43%) across tasks — evidence that the leverage is in amplifying people, not replacing them. Later reports track a widening AI skills gap.

The research call for agency

  • The Anthropic Institute research agenda

    Anthropic

    Names “designing interfaces that improve and promote human agency” as an open problem — alongside “group epistemology when populations use the same models” (homogenization) and “worker agency preservation.” coocBoard is a direct answer.

coocBoard is a personal engine that turns commodity compute into personal advantage — it layers empathy and your context onto raw capability, filters the world’s noise to what you need, and nudges you up the learning → doing → scaling arc toward what only you can build.