WritingReading path

Trust & Accountability.

Who is answerable when an agent decides. Accountability, provenance, fraud, and governance — the constraints that matter once meaning itself becomes cheap to produce.

12 essays in this path

of 94 published · 2023–2026

Signature figure
A stack of 12 signed documents, one per essay, the top one countersigned and stamped Is Apple harvesting adtech data? Do We Still Need Standards When Machines Understand Context? The First AI-Native Fraud Isn't a Fake — It's Laundered Authority Recommendation or Decision? The Veto That Doesn't Scale What Agents Can't Manufacture Whoever Writes the Constraints Owns the Outcome Measurement, on the Browser's Terms The Context Economy Has a Governance Problem The One Standard That Clears the Deal The Answerer of Record The Gate That Said No The Mandate Finished Last ACCOUNTED 12 essays
Accountability is the scarce thing once meaning goes to zero.
WHO OWNS THE AGENT'S DECISION The agent's decision — selection, timing, spend. Every choice it makes is downstream of the action space it was given: the agent chooses within the space, but it never chose the space. Agent decision selection · timing · spend The gate: the constraints the agent acts inside — goals, budget rails, guardrails — written and versioned by the agent owner. Constraint authorship is where the decision was really made; a bad outcome inside correctly enforced rails traces back to the rails. AUTHORED BY THE OWNER goals · budget rails · guardrails rules v0.4.2 The outcome — owned where the constraints were authored. The vendor owes enforceable rails and honest measurement; the user owes informed approval; ownership concentrates on the constraint author. Outcome owned where authored The append-only decision ledger: every consequential action recorded with the rules version in force. In the rules and still bad — the owner's constraints. Outside the rules — the vendor's enforcement. Approved over a warning — the user's approval. Fault localizes instead of diffusing. DECISION LEDGER · APPEND-ONLY #0417 · rules v0.4.2 · in-bounds · logged#0418 · rules v0.4.2 · in-bounds · logged#0419 · rules v0.5.0 · blocked · escalated not a feeling — a routing table accountability follows the action space Whoever writes the constraints owns the outcome.
Start here · Jul 2026Signature figure Canonical first-read

Whoever Writes the Constraints Owns the Outcome

Friday's poll asked who should be accountable for an agent's decision, and the room put half its ballots on The Agent Owner, with Shared Responsibility a distant second. This essay takes that answer seriously and then sharpens it: shared responsibility is what accountability looks like before a contract assigns it. In practice, accountability follows the action space — the vendor owes enforceable rails and honest measurement, the user owes informed approval, but ownership concentrates on whoever authors the constraints the agent acts inside. Which is why my vote went to The Agent Owner, and why the fix is infrastructure, not sentiment: versioned operating rules plus an append-only decision ledger turn 'shared' from a shrug into an audit trail.

Every failure mode routes to a party, because the routing was built before the failure.
Evgeny Popov
All essays

More on this path.

11 essays · newest first

2026 · 10 essays

  1. The Mandate Finished Last

    Season 2 opened by proposing that once every advertiser has a capable buying agent, advantage moves to the quality of the mandate. Forty-two operators voted, 78% chose proprietary data, and the mandate came last at 4%. The room is right about the market as it stands, and the reason has nothing to do with which asset actually decides the outcome.

  2. The Gate That Said No

    I rebuilt the retrieval system behind Ask Evgeny in one day: traces on every query, a 975-node knowledge graph, authority buckets on every position, and three model upgrades running in shadow. The most important thing that shipped is the feature that measured itself and refused to launch.

  3. The Answerer of Record

    The Season 1 finale. Eight Friday polls were one question asked eight ways — who issues the credential that lets an agent act? — and the room's own ballots supply eight structural constraints on the answer. Run the three candidates against all eight and only one survives: protocols carry the credential, constitutions bound it, contracts create it. The issuing layer is an Answerer of Record — the agentic successor to adtech's seller-of-record — because trust is conferred, never computed.

  4. The One Standard That Clears the Deal

    Season 1's last poll asks which standard matters most in five years: Identity, Measurement, Agent Protocols, or Trust & Accountability. This essay's answer: three of the four only reduce friction — a market can run without them, just more slowly and at more risk. Only one clears or blocks the transaction outright. That's the test that actually separates a standard from infrastructure, and it's independent of which option the room eventually picks.

  5. The Context Economy Has a Governance Problem

    An AdExchanger column just named the shift the whole industry is feeling: from an audience economy to a context economy. It is right, and it is the direction a season of this site's writing has been pointing. But 'understand the moment and act on it instantly' is one governance layer away from being a faster surveillance economy — and that layer is the whole ballgame.

  6. Measurement, on the Browser's Terms

    The W3C Attribution API debate, read as a constitutional question. Mozilla's Martin Thomson says the quiet part precisely: the test of any privacy technology is not the information flows — it is who decides. The browser is bidding to become the counterparty that grants measurement its standing, and every operator concern in the spec follows from that one move.

  7. What Agents Can't Manufacture

    Friday's poll asked what becomes scarce in a market run by agents. The obvious answer is the last decade's winner — the biggest data. Wrong layer. Every cycle relocates scarcity rather than abolishing it, and an agentic market makes data, compute, and distribution all self-suppliable — leaving one asset a machine is structurally incapable of minting for itself: trust.

  8. Recommendation or Decision? The Veto That Doesn't Scale

    Friday's poll asked when an AI recommendation becomes an AI decision. Most said 'when humans stop reviewing.' Right answer, wrong tense — in agentic ad tech that line is already behind us, and the surviving signature isn't a decision-maker, it's a name for the post-mortem.

  9. The First AI-Native Fraud Isn't a Fake — It's Laundered Authority

    Friday's poll asked what the first truly AI-native fraud is. The crowd answered with familiar villains. The trap was the question: the new scarcity isn't authenticity, it's accountability.

  10. Do We Still Need Standards When Machines Understand Context?

    The crowd voted 'more than ever' — but the real question isn't whether AI kills standards. It's which standards survive when meaning goes to zero and accountability becomes the scarce thing.

2023 · 1 essay

  1. Is Apple harvesting adtech data?

    The strategic read on App Tracking Transparency — Apple's privacy posture as competitive intelligence.

Other reading paths

Adjacent threads.

Every path

Or browse all writing →