A01 THE COMPLETE INDEX
THE ARCHIVE
Thirty-five notes across twenty-four months. Sorted by the period examined, published as a transparent retrospective in September 2026.
Archive dates describe the period discussed. They are not original publication dates. Every note carries its real publication status and source ledger.
35 NOTES SHOWN
The Constraint Is the Product
Frontier markets become legible when constraints stop being footnotes and become product inputs.
Why Frontier Markets First Look Like Coordination Problems
Supply exists. Demand exists. Talent exists. The missing product is often coordination.
Science Has a Translation Problem, Not an Idea Problem
The distance between an interesting mechanism and a useful intervention is an operating system.
The Difference Between Compute Capacity and Usable Capacity
GPU count is an input. The customer buys a workload that runs.
Tokenizing Science Is Easy. Coordinating It Is Hard.
A token can represent participation or rights. It cannot, by itself, run an experiment.
A GPU Is a Financial Asset Wearing a Software Interface
Behind every clean API sits capex, depreciation, utilization, debt service and customer credit.
Prediction Is Not Proof
The first answer is where scientific work begins, not where it ends.
Why Power Is the First Product in AI
Every token begins upstream, with electrons that can reach the right rack at the right time.
The Trust Stack for Open Science
Openness works when identity, provenance, permissions and accountability are designed together.
What an AI Scientist Should Be Allowed to Do
Delegate by reversibility, evidence quality and the cost of being wrong.
From Papers to Projects: The Missing Collaboration Layer
Knowledge becomes work only when people, permissions, files, decisions and incentives share a surface.
The Economics of a GPU Hour
The cheapest hour is not always the one that completes the workload at the lowest cost.
Why Scientific Benchmarks Miss the Real Work
Research is a continuous workflow. Most benchmarks inspect isolated moments.
IP-Sensitive Collaboration in an Open World
Teams want feedback and visibility without donating the asset they are trying to build.
The Unsexy Moat: Commissioning, Uptime, and Recovery
The glamorous part is the GPU. The moat is the 2 a.m. recovery process.
Evidence Before Claims: A Better Consumer Health Playbook
Mechanism, ingredient evidence and finished-product outcomes are different claims.
Why Bounties Sound Wrong to Biotech Operators
The mechanism may be useful. The language can still signal the wrong operating model.
The Data Center Has Become Part of the Product
For production AI, geography, cooling, network and uptime reach all the way into user experience.
From Scientific Community to Product Feedback Loop
Communities can connect research, product use and new questions, if data rights remain clear.
The Capital Stack Behind Frontier Progress
Different uncertainties deserve different capital, terms and time horizons.
Why Marketplaces Fail When Trust Is Scarce
Aggregation without verification multiplies listings faster than it multiplies fulfillment.
Milestones, Not Announcements
Proposed, funded, contracted, built, validated and operating are different realities.
What 90% Accuracy Does Not Tell You
The score is a beginning. The error distribution is the operating manual.
The Case for Researcher-in-the-Loop AI
The best agent does not remove judgment. It makes judgment more leveraged and inspectable.
Grid Queues Are Product Roadmaps
Interconnection timing determines which compute promises can become customer offers.
How Null Results Build Brands
A company that promises to publish the result, not its direction, is making a different kind of claim.
The Science-to-Shelf Flywheel
Research can inform products, and responsible product use can fund better questions.
Sovereign AI Is an Infrastructure Requirement
Control over models is incomplete without control over data, capacity, geography and operations.
The Open/Private Boundary in Collaborative Science
The boundary should be an explicit product surface, not an implicit social norm.
Why AI Infrastructure Sales Are Not SaaS
The sale combines technical fit, project delivery, asset finance and customer operations.
From Agreement to Workload: The Real Deployment Timeline
Commercial scope, engineering, integration, commissioning and acceptance are one continuous product.
A Protocol Is More Defensible Than an Ingredient List
Formulas are copyable. A disciplined system of selection, measurement and reporting is harder to reproduce.
Who Pays for Scientific Coordination?
The work between discovery and experiment is real work, even when no institution budgets for it.
Inference Turns Infrastructure into Customer Experience
When the model is the product, latency, uptime and cost per token become product metrics.
The Same Operator Problem in Biotech and Compute
Convert uncertain supply, difficult coordination and complex evidence into systems that people can trust.