More from Vision | TinyFish

More from Vision | TinyFish

Vision

CURRENT

A journal for living in the agentic age

Practitioner's Corner

Monday, August 24

ISSUE #43

Vision

One finding in this issue stopped us. Doctors who worked alongside an AI detection tool got measurably worse at spotting problems without it. Routine cases are where competence stays sharp, and the AI had taken the routine away. We kept running into that same gap in different forms. Agent memory drops provenance between sessions, so the system trusts recalled knowledge it can no longer trace. Approval gates land after the point where a person could still change the outcome. The checkpoints look right. Whether anyone behind them can still do the job is the question running through this whole issue.

Monday, August 24

ISSUE #43

Vision

One finding in this issue stopped us. Doctors who worked alongside an AI detection tool got measurably worse at spotting problems without it. Routine cases are where competence stays sharp, and the AI had taken the routine away. We kept running into that same gap in different forms. Agent memory drops provenance between sessions, so the system trusts recalled knowledge it can no longer trace. Approval gates land after the point where a person could still change the outcome. The checkpoints look right. Whether anyone behind them can still do the job is the question running through this whole issue.

Past Articles

August — Issue #43

Who Budgets for the Cases That Don't Fit

Signal and Status

Who Budgets for the Cases That Don't Fit

Organizations consistently underresource exception handling, treating their most consequential remaining human work as residual cost.

What Exceptions Know

Signal and Status

What Exceptions Know

Exception cases carry signals about shifting reality that dashboards cannot express, and automation may be closing that feedback channel.

The Exception Economy

The Exception Economy

The Exception Economy

When AI absorbs routine work, exceptions concentrate both cost and learning in the same place, creating a tension organizations haven't faced.

August — Issue #42

The Balance Nobody Checks

Paired Applications

The Balance Nobody Checks

Organizations automate exceptions into rules without tracking the judgment spent to create them, depleting a resource nobody accounts for.

The Price Tag Is the Wrong Number

Paired Applications

The Price Tag Is the Wrong Number

Why dollar value is the wrong axis for agent delegation when cancellation windows and information propagation determine real cost.

What Rollback Doesn't Return

What Rollback Doesn't Return

Agent risk has less to do with dollar amounts than with whether an action's consequences survive the undo button.

August — Issue #41

Nothing Was Down and Everything Was Wrong

Power and Fidelity

Nothing Was Down and Everything Was Wrong

An agent can clear every authorization check and still choose badly for you. Fidelity has no instrument that survives a real transaction, and the buyer eats the difference.

Somebody Else Decides Who Gets Asked

Power and Fidelity

Somebody Else Decides Who Gets Asked

Publishing structured data costs you nothing but effort. Getting invoked is somebody else's decision — and the leverage may sit with the thinnest registry, not the richest protocol.

Nobody Owns the Feed

The Visibility Shift

Nobody Owns the Feed

Something is reading your site literally now. When declared data becomes a commitment, drifting feeds, drip pricing and value that only shows up in conversation all start to vanish from view.

August — Issue #40

The Learning You Never Paid For

Skill and Stakes

The Learning You Never Paid For

Organizations got human development free, as a byproduct of productive work. When agents remove the middle of the task, that subsidy quietly lapses and nothing on the dashboard says so.

What the Diff Doesn't Keep

Skill and Stakes

What the Diff Doesn't Keep

Reviewing what a machine made will catch flaws. The judgment to imagine a better option appears to be imported from having built things yourself.

Nobody Counts the Boring Pages

Expertise Pipeline

Nobody Counts the Boring Pages

AI is eating the boring work that used to build engineers. The evidence points at the pipeline, not at the experts already in it.

July — Issue #39

The Counterparty Reflex

Two Sides of Acceleration

The Counterparty Reflex

Websites, payment networks, and content owners are building recognition infrastructure that will define agent access on their terms, before agents earn trust rather than after.

Speed Without Contestability

Two Sides of Acceleration

Speed Without Contestability

Agents act faster than organizations can challenge their output, degrading the review layer that makes clerk-model deployment survivable.

When Damage Outruns Detection

The Speed Variable

When Damage Outruns Detection

Agent speed turns familiar errors into cascades that outrun an organization's capacity to detect, absorb, or correct them. The binding constraint on adoption is not production but reception.

July — Issue #38

Review-Shaped Work

Evidence and Roles

Review-Shaped Work

When agents handle routine work, human responsibility doesn't diminish. It concentrates into higher-stakes review that demands context, authority, and genuine organizational redesign. Without those, oversight becomes decorative.

The Clerk Pattern

Evidence and Roles

The Clerk Pattern

Production agents across security, coding, and commerce keep converging on one architecture: scoped authority producing artifacts for human review, not autonomous action. The pattern works, until the interaction itself is the product.

The Clerk, Not the Copilot

The Clerk, Not the Copilot

AI agents aren't maturing toward peer-like collaboration. They're gaining institutional admissibility as clerks, and the price of admission is forcing organizations to make rules explicit that should have been explicit all along.

July — Issue #37

The One Thing You Can't Repurpose

Transfer and Gap

The One Thing You Can't Repurpose

Continuous delegation management has no organizational precedent, and evidence shows why existing access control can't substitute for it.

The Agent Governance Capabilities You Already Have

Transfer and Gap

The Agent Governance Capabilities You Already Have

Regulated industries already maintain compliance, QA, and exception-handling capabilities that map onto agent governance requirements in surprising ways.

The Delegation Muscle

The Delegation Muscle

Organizations that already codified delegation, authorization, and dispute recovery are absorbing agents fastest because they've already solved the hard problems.

July — Issue #36

Green Dashboards, Missing Answers

Where Standing Exists / Where It Doesn't

Green Dashboards, Missing Answers

Enterprise agents execute successfully but lack institutional standing, leaving organizations with green dashboards and no clean answers.

Why Payments Are Building Agent Standing First

Where Standing Exists / Where It Doesn't

Why Payments Are Building Agent Standing First

How chargebacks and disputes are forcing payment networks to build the accountability architecture agents need before anyone else.

The Standing Problem

Enterprises can use AI but can't scale agents because successful actions still lack the institutional legitimacy organizations need to own them.

June — Issue #35

The Demo-Mode Trap

Accountability in Practice

The Demo-Mode Trap

Without costly consequences forcing delegation clarity, low-stakes domains keep agents stuck as fast interns producing unaccountable output.

Why the Most Regulated Industries Are Deploying Agents First

Accountability in Practice

Why the Most Regulated Industries Are Deploying Agents First

Payments and compliance already built the scaffolding agents need, making regulated industries surprising frontrunners in agent deployment.

The Accountability Inversion

The Accountability Inversion

Agent adoption tracks accountability infrastructure, not model capability. The environments that can already answer who authorized what, within what limits, and where the records live are the ones reaching production first.

June — Issue #34

The Conditions for Meaning

Individual Shift / Organizational Conditions

The Conditions for Meaning

Meaningful AI-augmented work depends on specific design choices most organizations are making by accident, with compounding consequences.

The Taste You Didn't Earn

Individual Shift / Organizational Conditions

The Taste You Didn't Earn

Reviewing AI outputs builds real pattern recognition, but it may be a different thing than the domain expertise it resembles.

What Smoothness Costs

The Texture Problem

What Smoothness Costs

Organizations automating knowledge work are stripping out the friction where judgment forms, and most don't realize it until the expertise is gone.

June — Issue #33

Monitoring the Monitors

Lessons and Designs

Monitoring the Monitors

What would an early warning system for degrading human oversight actually look like, and why is it so hard to build?

What Aviation Built After It Noticed the Problem

Lessons and Designs

What Aviation Built After It Noticed the Problem

Aviation built concrete measurement infrastructure before its automation countermeasures. Enterprise AI skipped that step entirely.

The Core Problem

Smooth AI assistance decouples confidence from accuracy at the individual level, then compounds into organizational blind spots that existing oversight cannot detect.

June — Issue #32

What the Six Percent Are Doing Differently

Two Consequences

What the Six Percent Are Doing Differently

The few organizations redesigning work around agents share a method: decompose workflows first, then assign decision rights deliberately.

Where Does Judgment Come From Now?

Two Consequences

Where Does Judgment Come From Now?

As agents absorb the routine tasks that built professional intuition, the pipeline for developing judgment quietly thins.

The Boring Turn

The Boring Future

The Boring Turn

Enterprise AI agents are crossing from exciting to invisible, and the hardest questions only surface after the technology stops being the conversation.

June — Issue #31

When Smooth Work Feels Like Good Work

Window and Mechanism

When Smooth Work Feels Like Good Work

AI makes work feel fluent, and humans read fluency as accuracy. Designing around this cognitive mechanism matters more than adding checkpoints.

What the Maturity Gap Is Actually For

Window and Mechanism

What the Maturity Gap Is Actually For

The gap between AI piloting and autonomy is a fleeting window to build evaluation infrastructure before operational knowledge disappears.

The Expertise You're Losing Is the Expertise You'd Need to Notice

Expertise Decay

The Expertise You're Losing Is the Expertise You'd Need to Notice

Organizations lose the capacity to detect expertise erosion on the same timeline they lose the expertise itself, and the recursion is already underway.

May — Issue #30

The Discovery Problem No One Has Solved

Two Consequences

The Discovery Problem No One Has Solved

Agents can talk to each other but cannot find each other, and history shows discovery layers rarely stay neutral without deliberate design.

Agent-First Design and the Fate of Human-Experience Moats

Two Consequences

Agent-First Design and the Fate of Human-Experience Moats

When agents evaluate services at machine speed, competitive advantage migrates to invisible properties no marketing team has learned to measure.

The Indifferent Buyer

The Inverted Audience

The Indifferent Buyer

When the buyer has no eyes or impulses, the web's trillion-dollar persuasion economy loses its audience and commerce rewires around machine legibility.

May — Issue #29

What the Junior Analyst Knew

Two Defaults Hiding in Plain Sight

What the Junior Analyst Knew

Automating entry-level work doesn't just thin the talent pipeline—it removes the institutional feedback loop organizations never knew they depended on.

The Pilot That Worked

Two Defaults Hiding in Plain Sight

The Pilot That Worked

Successful AI pilots quietly encode current workflows as permanent infrastructure, closing the window for organizational redesign before anyone notices.

The Architecture of Foregone Conclusions

The Anchor Argument

The Architecture of Foregone Conclusions

Temporary infrastructure choices are calcifying into permanent architecture faster than governance can respond, and the window for intentional design is closing.

May — Issue #28

Why Oversight Feels Like Loss

Why Oversight Feels Like Loss

Shifting from doing work to overseeing AI replicates a familiar loss of craft identity that better training alone cannot resolve.

Where Meaning Hides Inside Work

Where Meaning Hides Inside Work

Workers' resistance to automating enjoyable tasks reveals where meaning lives in work and where agent deployment will meet friction.

The Signal in the Satisfaction Gap

The Paradox Framed

The Signal in the Satisfaction Gap

Rising productivity plus falling engagement signals a work design problem enterprises must solve before agent orchestration roles are built by default.

May — Issue #27

The Platform Without an Audience

What Breaks Next

The Platform Without an Audience

When your customer is software, ads, loyalty programs, and storefronts lose their audience. What's left?

The Liability Vacuum

What Breaks Next

The Liability Vacuum

Agents can spend money but can't be sued. The accountability framework is still a draft while commerce rolls forward.

The Scaffolding Nobody Voted On

The Core Reframe

The Scaffolding Nobody Voted On

Agents are quietly gaining wallets, identities, and credit limits while nobody debates what economic participation means without wanting.

April — Issue #26

How to Divide What You Haven't Done

Two Implications

How to Divide What You Haven't Done

Breaking complex work into agent-sized pieces requires deep domain intuition that automation itself is quietly eroding.

The Foreman Problem

Two Implications

The Foreman Problem

Supervising agents that act autonomously demands a cognitive discipline we haven't named, trained for, or even fully recognized.

The Consequences Nobody Designed For

The Core Parallel

The Consequences Nobody Designed For

Assembly lines split craft from quality control a century ago—multi-agent AI is doing the same to knowledge work now.

April — Issue #25

Agents Don't Browse

Agents Don't Browse

The SEO playbook assumed browsers who wander. Agents that execute purchases break the analogy where it matters most.

The Web Is Learning to Talk Back

The Web Is Learning to Talk Back

From static text files to live queryable endpoints, websites are quietly rebuilding themselves to serve non-human visitors.

When the Web's Visitors Start Buying

The Second Front Door

When the Web's Visitors Start Buying

The web built infrastructure for visitors that read. Its new visitors spend.

April — Issue #24

How Do You Sell Something That Succeeds by Vanishing?

How Do You Sell Something That Succeeds by Vanishing?

Agents succeed by disappearing, but brands need visibility. The commercial paradox reshaping pricing, bundling, and loyalty.

The 600 Agents Nobody Hired

The 600 Agents Nobody Hired

When the most valuable agents are the ones nobody notices, oversight loses the friction it always depended on.

When the Thing That Disappears Is the Thing That Decides

When the Thing That Disappears Is the Thing That Decides

Agents are sliding from branded products to invisible infrastructure, but invisible judgment isn't invisible plumbing—and governance always arrives late.

April — Issue #23

PwC's Hourglass Org Runs on a Reservoir It Can't Refill

Two Bets on What Replaces It

PwC's Hourglass Org Runs on a Reservoir It Can't Refill

PwC's generalist workforce model runs on deep specialists. What happens when nobody goes deep first?

IBM Bets That Watching Builds Knowing

Two Bets on What Replaces It

IBM Bets That Watching Builds Knowing

Can reviewing AI output build the same expertise as years of hands-on failure? IBM's hiring surge tests that bet.

The Infrastructure Nobody Budgeted For

The Missing Rung

The Infrastructure Nobody Budgeted For

Organizations are quietly draining the reservoir that produces experienced practitioners, and they won't notice until it's empty.

March — Issue #22

The Pressure Gradient

Two Faces of It

The Pressure Gradient

Business units and security teams aren't misaligned on AI agents—their reward structures are.

Who Answers for the Agent

Two Faces of It

Who Answers for the Agent

When agents act without judgment, accountability lands on a role that doesn't exist yet.

The Org Chart Nobody Drew

The Structural Problem

The Org Chart Nobody Drew

Enterprises aren't just deploying agents without oversight—they're accidentally building a parallel workforce faster than anyone can govern it.

March — Issue #21

Who Decides What an Agent Can Trust?

New Rules, New Gatekeepers

Who Decides What an Agent Can Trust?

Agent commerce needs machine-readable trust, and whoever defines it becomes the infrastructure nobody can remove.

When the Customer Has No Eyes

New Rules, New Gatekeepers

When the Customer Has No Eyes

When AI agents replace human browsing, businesses that can't be parsed don't get penalized—they vanish entirely.

What Breaks When the Buyer Has No Brain to Hack

The Persuasion Stack

What Breaks When the Buyer Has No Brain to Hack

The commercial web is a persuasion machine disguised as an information network, and agents are calling the bluff.

March — Issue #20

The Maintenance Phase

Two Consequences

The Maintenance Phase

Seventy percent of successful AI investment goes to people, not technology. The unglamorous work decides everything.

The Feature, Not the Product

Two Consequences

The Feature, Not the Product

Agents are being absorbed into platforms like databases before them. The value is migrating to infrastructure.

The Fastest Boring in Tech History

The Boring Inflection

The Fastest Boring in Tech History

Agents are becoming infrastructure at unprecedented speed, but organizations may be absorbing the technology faster than they understand it.

March — Issue #19

What Stops Being Found When Every Agent Agrees

Default Settings

What Stops Being Found When Every Agent Agrees

When every agent optimizes the same way, the web's weird, wonderful long tail quietly stops being discoverable at all.

What Counts as Good When Agents Decide

Default Settings

What Counts as Good When Agents Decide

AI shopping agents don't just filter products—their opaque criteria quietly harden into the operative definition of quality itself.

The Rise of Ghost Institutions

The Rise of Ghost Institutions

AI agents are encoding institutional taste at massive scale, and nobody involved in building them recognizes they're making aesthetic judgments.

March — Issue #18

Secondary interfaces always decay. What happens when humans prefer the agent's door because it wasn't designed against them?

The Honest Interface

The Honest Interface

Dark patterns can't survive serialization. When agents replace visual interfaces, manipulation loses its medium.

The Honest Version We Built for Machines

The Honest Version We Built for Machines

Building an honest interface for agents accidentally exposes what the human-facing web was designed to hide all along.

February — Issue #17

When Judgment Stops Being Scarce

The Professional Services Threshold

When Judgment Stops Being Scarce

Professional judgment is splitting into pattern-matching that scales infinitely and accountability that never will, rewriting how firms exist.

Why Anthropic Engineers Spent Six Months Inside Goldman

Proving Ground / New Model

Why Anthropic Engineers Spent Six Months Inside Goldman

Six months of embedded engineers signals the infrastructure gap between model capability and production reliability in high-stakes environments.

Goldman's Compliance Bet Signals Where AI Agents Actually Are

Proving Ground / New Model

Goldman's Compliance Bet Signals Where AI Agents Actually Are

Domain selection reveals capability maturity—Goldman chose compliance because of its zero-error tolerance and regulatory stakes.

February — Issue #16

Why Cost Predictability Became a Reliability Problem

Infrastructure Economics

Why Cost Predictability Became a Reliability Problem

Cost volatility created a reliability problem when vendor pricing swings generated operational risk that uptime percentages couldn't measure or governance couldn't prevent.

When Continuous Workloads Break Cloud Economics

Infrastructure Economics

When Continuous Workloads Break Cloud Economics

Continuous inference workloads exposed limits in cloud economics built for episodic spikes, forcing infrastructure decisions around predictable costs over elastic capacity.

The Morning Check That Stopped Happening

Threshold Moments

The Morning Check That Stopped Happening

When daily verification rituals quietly dissolve into trust, workers discover their expertise has relocated from managing tools to evaluating output—a threshold crossed through accumulated irrelevance rather than conscious decision.

February — Issue #15

The Years Your Organization Spends Escaping Platform Decisions

Infrastructure Economics

The Years Your Organization Spends Escaping Platform Decisions

Platform migrations consume years of organizational capacity—the real cost isn't the invoice, it's what you can't build while executing the move.

When Power Availability Dictates Where You Can Build

Infrastructure Economics

When Power Availability Dictates Where You Can Build

Grid constraints override traditional location economics—power availability within deployment timelines now dictates where infrastructure can actually be built.

When Execution Speed Stops Being the Constraint

Threshold Moments

When Execution Speed Stops Being the Constraint

When typing code manually triggers visceral irritation, organizations discover execution speed has stopped constraining output—and team structures designed around slower work can't reverse course.

February — Issue #14

The Geographic Necessity Premium

Infrastructure Economics

The Geographic Necessity Premium

Regional presence becomes mandatory when sites detect geographic anomalies, creating necessity costs that infrastructure optimization strategies cannot eliminate.

When Cheaper Infrastructure Creates Higher Costs

Infrastructure Economics

When Cheaper Infrastructure Creates Higher Costs

Infrastructure arbitrage that saves on compute creates operational losses when sites resist automation differently by region at production scale.

What Middle Management Actually Built

Threshold Moments

What Middle Management Actually Built

Organizations are flattening management layers faster than they're rebuilding the judgment-development infrastructure those layers quietly provided—revealing what middle managers actually built beyond reports.

January — Issue #13

Paying for Capabilities You Haven't Needed Yet

Infrastructure Economics

Paying for Capabilities You Haven't Needed Yet

The flexibility premium buys organizational capacity to handle web complexity you haven't encountered yet—insurance against an adversarial environment that constantly evolves.

When Infrastructure Becomes Irreversible

Infrastructure Economics

When Infrastructure Becomes Irreversible

Infrastructure decisions become permanent through organizational learning, not technical constraints—teams accumulate expertise that doesn't transfer when you try to migrate.

When Organizational Capacity Becomes the Binding Constraint

Threshold Moments

When Organizational Capacity Becomes the Binding Constraint

Organizations discover their readiness gaps only during production deployment—when infrastructure complexity, governance frameworks, and cross-system coordination reveal what pilots never exposed.

January — Issue #12

The Legitimacy Premium in Emerging Authentication Standards

Infrastructure Economics

The Legitimacy Premium in Emerging Authentication Standards

Emerging cryptographic standards create arbitrage windows where early authentication investments become strategic positioning as verified identity turns scarce.

The Authentication Paradox at Production Scale

Infrastructure Economics

The Authentication Paradox at Production Scale

Authentication complexity at scale inverts from infrastructure tax to competitive moat as reliability gaps compound across thousands of sites.

When "Working" Stops Meaning What It Used To

Threshold Moments

When "Working" Stops Meaning What It Used To

When your extraction dashboard shows green but the data's semantically wrong, you've crossed into probabilistic infrastructure—where "working" means something fundamentally different and your entire validation framework must evolve.

January — Issue #11

How Infrastructure Waste Compounds Over Time

Infrastructure Economics

How Infrastructure Waste Compounds Over Time

Infrastructure waste from consolidation doesn't just persist—it compounds over years through migration costs, shadow systems, and embedded inefficiency that exceeds initial savings.

What Consolidation Actually Costs at Production Scale

Infrastructure Economics

What Consolidation Actually Costs at Production Scale

Consolidation promises savings but creates infrastructure waste through over-provisioning, monitoring multiplication, and error recovery overhead that compounds with organizational complexity.

When Checking Costs More Than Running

Threshold Moments

When Checking Costs More Than Running

At scale, the infrastructure checking your web agents breaks more often than the agents themselves—forcing teams to abandon comprehensive verification for risk-based confidence.

January — Issue #10

When Scale Inverts Economics

Infrastructure Economics

When Scale Inverts Economics

Adversarial web environments invert traditional infrastructure patterns where optimization multiplies costs and scale increases unit economics through detection complexity.