Sign inGet started
Journal

Thinking out loud

On AI skills, the creator economy, and what happens when expert judgment becomes portable.

Knowledge Work15 Mar 2026 · 5 min

The Judgment Gap — Why AI Knows Everything and Understands Nothing

AI can access the world's knowledge in milliseconds. But knowledge without judgment is just noise. The next wave of AI value isn't better models — it's better methodology.

Read article
Platform

A Textbook vs. The Professor — Why the Future of AI Expertise Is Person-Based, Not Framework-Based

A skill is a textbook — one framework, applied well. An agent is the professor — all frameworks, interconnected, with the judgment to route between them. Nobody asks "apply the methodology." They ask "what would the expert think?"

8 Mar 2026 · 10 min
Knowledge Work

The End of Generic AI Advice

Ask base AI to review your cold email and you get generic advice about personalisation. Run it through an expert skill and you get a calibrated score, an offer tier diagnosis, and a rewritten version using pattern-interrupt psychology. Same model. Radically different output.

1 Mar 2026 · 10 min
Creator Economy

Your Expertise Is Being Consumed by AI. You're Not Getting Paid.

Every expert's published methodology is now part of the training data that makes base AI "know about" their framework. But "knowing about" isn't "being able to apply." The value chain doesn't compensate creators. There's an alternative.

22 Feb 2026 · 7 min
Skills

Download Expertise, Not Just Information

Not a chatbot, not a course, not a consultant — downloadable expertise. What happens when proven methodology applies itself to your actual work, adapts to your context, and produces artifacts you can use immediately.

15 Feb 2026 · 5 min
Creator Economy

What It Actually Looks Like to Build a Skill (And Where It Leads)

A transparent walk-through that follows a methodology from candidate to published skill — and then to the agent tier. Real pipeline data, real outputs, real economics. The article they read when they're ready to commit.

8 Feb 2026 · 9 min
Creator Economy

The Expert's New Revenue Model — From Hourly Billing to Methodology Licensing

A consultant bills $500/hour serving one client. The same methodology, packaged as a skill, serves thousands simultaneously. But the real revenue explosion comes at the Agent Tier: 10-50x the invocation volume.

1 Feb 2026 · 7 min
Platform

Skills Are the Path. The Agent Is the Destination.

Build your first skill from one book. The pipeline extracts thresholds, decision trees, patterns. Then build a second. At two or more, the platform offers Agent Unification. The agent is the destination — start with one skill.

25 Jan 2026 · 6 min
Creator Economy

The Claim Model — Your Methodology Already Has a Page

We've already researched your methodology. Users are expressing interest. You can claim your skill page and work with us to make it real — or let it remain as an inspired placeholder.

18 Jan 2026 · 3 min
Creator Economy

Why a 70% Revenue Share (And Why That Matters)

The platform handles extraction, testing, hosting, delivery, IP protection, billing, analytics, and evolution. The creator brings the methodology — which is the entire value. 70% to the creator because the methodology IS the product.

11 Jan 2026 · 3 min
Technical

SKILL.md — The Open Standard for Packaging AI Expertise

Structured markdown files that package methodology for AI consumption. MCP is "USB-C for AI" — tool provisioning. SKILL.md is expertise packaging. Together they form the infrastructure layer for the next generation of AI applications.

4 Jan 2026 · 4 min
Platform

Person Over Position — Why Licensed Creator Agents Beat Generic Role Agents

Role agents encode generic capabilities anyone can replicate — 61 free role agents exist on GitHub right now. Person agents encode specific expert judgment that took decades to develop. The gap is our entire value proposition.

28 Dec 2025 · 4 min
Platform

How We Test Skills — Inside the Methodology Fidelity Pipeline

Methodology fidelity testing, conversation grading, journey testing across 30-50 turns, comparative testing across 12 personas, independent judge validation. Published test results are the trust signal the market needs.

21 Dec 2025 · 7 min
Technical

MCP Has 17,939 Servers. Less Than 5% Are Monetised.

Tool provisioning is solved. Expertise monetisation is wide open. Most MCP servers are developer tools. What's missing is expert methodology delivered through the same protocol.

14 Dec 2025 · 6 min
Technical

Why Your AI Agent Needs a Trust Score

An AI agent choosing between random MCP tools and a tested skill needs something equivalent to product reviews. Trust infrastructure is the unsolved problem in agent commerce. Whoever builds it captures the market.

7 Dec 2025 · 6 min
Knowledge Work

When Your AI Agent Needs to Call an Expert

The agent knows Shape Up exists as a concept. It can summarise the framework. But it can't apply calibrated thresholds. So the agent does something new: it calls a skill — tested methodology, delivered via API, consumed in seconds.

30 Nov 2025 · 7 min
Skills

How Enterprise Teams Are Deploying Expert Methodology at Scale

Your team of 50 PMs can't each have a personal coaching relationship with a methodology expert. But they can each have access to tested methodology, delivered through any AI interface, with usage analytics and budget controls.

23 Nov 2025 · 6 min
Skills

The Agent Skills Your AI Assistant Is Missing

Your coding agent has GitHub, Jira, and Figma tools. But when you ask "should I build this?" your agent has no product strategy skill. The next competitive advantage is the quality of methodology, not the model.

16 Nov 2025 · 6 min
Platform

What Would Gary V Say? — Inside the Knowledge Graph That Powers Creator Agents

Traces a real query through the system: question embedded, semantic search finds relevant knowledge nodes, graph traversal pulls connected beliefs, contradiction detection surfaces evolving positions, unified voice synthesis responds.

9 Nov 2025 · 8 min
Knowledge Work

75% of Knowledge Workers Already Use AI. Most of Them Are Guessing.

The gap between AI adoption and AI effectiveness is where skills live. Most AI use is unsupported — workers figuring it out alone, producing inconsistent results. What if the AI came pre-loaded with proven methodology?

2 Nov 2025 · 6 min
Knowledge Work

The Consulting Industry's $500/Hour Problem

Consultants were never really selling knowledge — they were selling judgment, relationships, and accountability. The judgment component can now be extracted and delivered at scale. This isn't disruption — it's specialisation.

26 Oct 2025 · 8 min
Platform

The Living Coach — Why Static Skills Are Just the Beginning

A static skill catches "my direct report keeps missing deadlines" but misses "my teammate isn't stepping up." The Living Coach catches both because it understands intent, not just keywords. Three intelligence layers that change everything.

19 Oct 2025 · 8 min
Skills

What "AI-First" Actually Means (It's Not a Chatbot)

The distinction between "AI-added" products and "AI-first" products. An AI-first expertise experience is something different: the methodology applies itself to your work, adapts to your context, and produces artifacts immediately.

12 Oct 2025 · 6 min
Skills

Designing for Resistance — What Coaching Psychology Teaches AI Product Teams

Most AI products assume users want to be helped. In practice, people resist good advice — especially when it challenges their assumptions. Three types of resistance and how skills are designed to navigate each.

5 Oct 2025 · 7 min
Skills

The Six Things an Expert Does That AI Doesn't (Yet)

Calibrated judgment. Sequenced diagnosis. Prescribed formats. Pattern recognition. Contextual memory. Tool orchestration. Each can be extracted from experts and delivered through AI — but only with the right infrastructure.

28 Sept 2025 · 6 min
Creator Economy

What Happens When AI Can "Do" Your Framework?

Base Claude can explain Shape Up. It can list the stages, define appetite, describe pitches. But when a real PM brings a real problem, base scores 26.6/40 while the skill scores 31.4/40. "AI can explain it" is not "AI can do it."

21 Sept 2025 · 6 min
Platform

Why We Built on Open Standards (MCP + SKILL.md)

If the format is open, what's the moat? The extraction pipeline, testing infrastructure, creator relationships, IP-protective delivery, and trust scores. The format is open. The methodology quality is proprietary.

14 Sept 2025 · 5 min
Platform

Introducing Universal Credits — Why We Killed Seat-Based Pricing

Seat-based pricing is declining, creates lower margins, and doesn't work for agents. Credits are fungible across skills, transparent, agent-native, and aligned with usage. One credit equals ten cents.

7 Sept 2025 · 5 min
Become a creator

Turn your expertise into an AI skill

Your methodology. Your judgment. Available to anyone, on demand.

Start with one skill →