Field notes · Study notes · Analysis

Notes from the work
around technology.

I write about what happens when a requirement reaches a product, a workflow and a person who must make the next decision. Explore field notes and analysis on systems, AI, IP and data.

Russian originals are adapted as English site editions, not translated word for word. Repeated platform versions are combined into one canonical article.

Featured study note

How Much Is an AI Agent Allowed to Spend?

Following one task through token accounting, P95 forecasts and the moment its budget runs out.

Read the site edition

Three routes through the library

Start with the question in front of you.

Recent writing

Recently added to the library.

The latest analysis, including AI task costs, forecasts and spending authority.

How Much Is an AI Agent Allowed to Spend?

Following one task through token accounting, P95 forecasts and the moment its budget runs out.

Read site edition

Do Not Rebuild the Company for Every External Question

A working design for connecting repeated external statements to current records, named owners, scope, version and review triggers.

Read on LinkedIn

Russia’s 2026 AI Law Is Not a General AI Act

Why Federal Law No. 243-FZ is better understood as a foundation-model sovereignty law, and what that leaves outside its primary frame.

Read on LinkedIn

Start here

Three pieces that explain the operating method.

One on the company around the product, one on AI in daily work and one on cross-market privacy architecture.

The Product Works. Is the Company Around It Ready?

A practical diagnostic for the moment when a growing product can no longer rely on memory, informal ownership and undocumented handoffs.

Read site edition

AI Is Already in the Workflow. What Changes Now?

AI rarely enters a company as one large programme. It arrives as a browser tab, a plugin or a familiar service with a new capability.

Read site edition

Build the Core, Localise the Delta

A privacy architecture can travel across markets only if the common operational layer is separated from the role-specific legal decision.

Read site edition

Systems · 6 pieces

Systems and Implementation

How a growing product becomes a company capability: provenance, data maps, operating mechanisms and decision evidence.

The Product Works. Is the Company Around It Ready?

A practical diagnostic for the moment when a growing product can no longer rely on memory, informal ownership and undocumented handoffs.

Read site edition

A Repository Is Not a Product History

How to reconstruct the people, agreements, third-party components, data and AI tools behind a growing product before due diligence sets the deadline.

Read site edition

How to Draw a Working Data Map for a Live Product

Start with one event, follow every copy and put decisions beside the route. A practical data map is an operating instrument, not a decorative diagram.

Read site edition

AI Made Information Cheap. Attention Is Still Expensive.

We can generate research, analysis and documents faster than we can absorb them. The next AI productivity skill may be knowing when to stop.

Read site edition

A Designed System Is Not Yet a Working System

A document can be completed in a week. A mechanism becomes real only where someone makes a decision and either passes through it or works around it.

Read site edition

Do Not Rebuild the Company for Every External Question

A working design for connecting repeated external statements to current records, named owners, scope, version and review triggers.

Read on LinkedIn

AI governance · 9 pieces

AI Governance and IP

Use-case visibility, human control, frontier capability shifts and the rights architecture behind AI products.

How Much Is an AI Agent Allowed to Spend?

Following one task through token accounting, P95 forecasts and the moment its budget runs out.

Read site edition

AI Is Already in the Workflow. What Changes Now?

AI rarely enters a company as one large programme. It arrives as a browser tab, a plugin or a familiar service with a new capability.

Read site edition

Shadow AI Is Already Inside the Company

A ban can close access to a service. It cannot show where AI already participates in work, what data it receives or which decisions now depend on it.

Read site edition

A Human Clicked Approve. Was the Decision Actually Reviewed?

The button records human presence. Control begins where disagreement can change the outcome.

Read site edition

Can AI Governance Keep Up with Frontier Capabilities?

Risk classification is only the beginning. Governance also needs to notice when the same system becomes more capable, autonomous and able to act.

Read site edition

Whose Model Is It After Fine-Tuning?

A claim that a company owns 'the model' hides a stack of assets, licences, trade secrets, data rights and unresolved legal questions.

Read site edition

Russia’s 2026 AI Law Is Not a General AI Act

Why Federal Law No. 243-FZ is better understood as a foundation-model sovereignty law, and what that leaves outside its primary frame.

Read on LinkedIn

AI Governance: What It Actually Means in a Real Company

Why implementation starts with the tools already in use, the data entering them and the ownership questions most frameworks leave abstract.

Read on LinkedIn

An AI Sings in Someone Else's Voice

Notes on synthetic voice, creative identity and the ownership questions underneath an AI-generated performance.

Read on LinkedIn

Privacy and markets · 5 pieces

Privacy and Markets

From legal requirement to product behaviour, market-specific controls, evidence and responsible ownership.

Build the Core, Localise the Delta

A privacy architecture can travel across markets only if the common operational layer is separated from the role-specific legal decision.

Read site edition

Mapping the Saudi PDPL into Business Operations

Study notes on processing activities records and the point where a legal structure becomes a live map of systems, vendors, retention and responsibility.

Read site edition

Mapping the Saudi PDPL: Data Disclosure and Restrictions

A disclosure is not just a line in a privacy notice. It is a movement event with a purpose, recipient, restriction, transfer question and audit trail.

Read site edition

Same Principles, Different Teeth

The privacy principles travel. Enforcement does not. A comparative reading of where similar legal grammar produces different operational priorities.

Read site edition

Saudi Arabia Is Not a Larger Version of Qatar

A note on treating each market as its own operating context rather than scaling assumptions from a neighbouring jurisdiction.

Read on LinkedIn

From analysis to architecture

The article makes the question visible. The implementation work gives it an owner, a control and an evidence trail.