Skip to main content
Thierry Gilgen

The Studio

The Library

The long memory of the Studio.

A growing collection of Field Notes on systems, artificial intelligence, sovereignty, infrastructure, trust, preparedness, and human agency.

80 volumes

A place to begin

Choose a question from your own work. Each reading offers a practical starting point.

  1. Who takes responsibility after the AI pilot?

    Who owns the AI system after the pilot? 5 min

    An ownership map for decision rights, monitoring, escalation and exit once real work depends on the system.

  2. What should a board ask about digital dependencies?

    Digital sovereignty: a checklist for Swiss boards 5 min

    Six layers, six pieces of evidence, and a decision the organisation can actually carry out.

  3. What can we do when a supplier fails?

    The first ten minutes of a supplier outage 3 min

    A small exercise to locate the delay between recognising a failure and being able to act.

MEMEX

The living Library

Each point is a published article. Threads connect related volumes and briefs to their full volumes. Select an article, then open a reviewed explanation to understand why two volumes connect.

80 articles · 308 connections

Continue to the shelves

Points belong to shelves A–H. Their colours identify subjects; select a volume to follow its connections.

Follow a thought.

Select a point or search for a volume above. Its connections offer another path into the Library.

Published articleEditorial connectionExplained connection

Inspired by the associative trails of Vannevar Bush, As We May Think (1945).

The question beneath the volumes

The Library repeatedly returns to one problem:

How do humans retain the capacity to understand, govern, contest, repair, and leave the systems on which they depend?

Across engineering, AI, infrastructure, economics, information, and public institutions, several conditions recur:

  • systems must be understandable enough to challenge;
  • responsibility must remain operational rather than ceremonial;
  • critical dependencies require credible alternatives;
  • intervention must remain faster than escalation;
  • incentives must not silently replace intentions;
  • human and institutional competence must be preserved;
  • evidence requires provenance;
  • important decisions must remain contestable;
  • and the capacity to change direction must survive success as well as failure.

These are recurring findings across the published volumes — not eternal laws.

Volume

Collection ASovereignty

13 volumes

A.01

Vol. 06 Digital sovereignty is not nationalismVol. 15 Sovereignty Is Not Self-SufficiencyVol. 16 Human Sovereignty in the Age of Machine IntelligenceVol. 17 Sovereignty Is Lost One Optimisation at a TimeVol. 21 Perhaps We Should Simply Speak About SovereigntyVol. 22 The Exit TestVol. 28 Human SovereigntyVol. 29 Proof of HumanVol. 36 Open Is Not SovereignVol. 38 One Agent, Many SovereigntiesVol. 44 Digital sovereignty: a checklist for Swiss boardsVol. 65 The Sovereignty GapVol. 72 Control Cannot Be Promised

Collection BArtificial Intelligence

18 volumes

B.01

Vol. 03 Why AI projects failVol. 05 The optimisation trapVol. 07 AI strategy without operational ownership is theatreVol. 26 The Calendar Catches UpVol. 27 The Last InterfaceVol. 32 The Unappointed OperatorVol. 45 Who owns the AI system after the pilot?Vol. 48 Access Is Not AuthorityVol. 49 The Decision LayerVol. 55 Rediscovery Is Not RegressionVol. 58 The Fear PremiumVol. 61 The Obedient MachineVol. 62 The Work Between the PromptsVol. 73 We Gave AI the KeysVol. 74 What You Get from AIVol. 76 The Intelligence We Keep InventingVol. 78 AI in Accounting: How to Tell Whether It Pays OffVol. 80 Building an AI-First Organisation

Collection CInfrastructure

6 volumes

Collection DPreparedness

5 volumes

Collection ESystems & Trust

14 volumes

E.01

Vol. 02 The Shinkansen principleVol. 04 Build systems people trustVol. 12 Deterministic Judgment in AI-Assisted Investment ResearchVol. 23 The Fiduciary MachineVol. 34 The Appointed DoorVol. 40 AI-First Is an Operating ModelVol. 41 Stop Building an AI Org Chart. Build a Workflow.Vol. 42 Grok Bot Starter PackVol. 47 Incident Ownership Without a PrincipalVol. 50 The Room Between Our AgentsVol. 51 The World Does Not ResetVol. 66 The Permission That Exposed a NationVol. 68 Access Is Not AgencyVol. 77 The Defender’s Window

Collection FExploration

3 volumes

Collection GEconomy & Power

9 volumes

Collection HInformation & Knowledge

12 volumes

H.01

Vol. 13 When Classification Becomes RealityVol. 14 The Attribution CrisisVol. 19 Who Owns Your National Conversation?Vol. 20 The End of the Referral EconomyVol. 24 The Last Human CorpusVol. 25 The Scan Is Not the BookVol. 35 What Libraries DefendVol. 37 The City That Became a HeadlineVol. 39 The Search Console That Does Not ExistVol. 67 The Word Is Not YoursVol. 71 Owning the Platform Is Not Owning the ConversationVol. 79 The Book Became a System

Select a spine to inspect the volume, then open the Field Note. The shelves grow with the collection.

Recently added to the collection

  1. B · Artificial Intelligence

    Building an AI-First Organisation

    A one-second answer, a monthly AI bill and the question worth bringing to capital

  2. H · Information & Knowledge

    The Book Became a System

    Six days, five centuries, and the work between an answer and its evidence

  3. B · Artificial Intelligence

    AI in Accounting: How to Tell Whether It Pays Off

    AI in accounting: why the exceptions decide the benefit

Browse volumes by publication date →

Read what comes next

Receive Thierry’s newsletter through Substack, or follow new Library volumes in your feed reader.

Subscribe on SubstackFollow via RSS (English)