Critique & decision-making — Field Notes

DESIGN SYSTEMS COLLECTIVE

LDLR DESIGN FIELD NOTES

The rule breaks before the agent does

Negreiros publishes a real ADR, three agent permissions, and the Safari bug that broke production twice while every rule was followed exactly as written.

BY GUILHERME NEGREIROS / AUG 27, 2026

Agents Don't Break the Rules. The Rules Break First.

The rule was applied exactly as written. The regression came back the same day it left.

Why it matters

— ADRs agents read before proposing anything: context, decision, rejected alternatives

— Three permissions in the repo: observe, generate, never overrule a ratified decision

— WebKit blocks var() inside :visited on purpose; the token never resolved

— A compliance audit removed the hex fallback, correctly, and reintroduced the bug

— Even the post-incident note drifted; records need review, not write-once trust

Read on Design Systems Collective →

Systems & primitives Critique & decision-making

DESIGNER FUND

LDLR DESIGN FIELD NOTES

Protect the search, not the output

Saarinen, Aufmann, and Kumar on using AI to widen design's search instead of shortening it — context gathering in, critique and copy deliberately out. Our first video field note.

BY KARRI SAARINEN, CHARLIE AUFMANN, AND ISHA KUMAR / AUG 12, 2026

Protecting the Thinking Behind Great Design | Inside Linear's Design Team

When you can do anything, what you choose not to do is far more important than doing everything much faster.

Why it matters

— Karri's /linear-way skill interrogates 40k customer requests for the real problem

— The wiki ask, reframed: let context live where execution lives

— Kumar caps AI at 20–30% of net-new design work, on purpose

— Aufmann brings raw customer quotes to review, so debate runs on evidence

— Define what good looks like first — otherwise the output can never fail

Read on Designer Fund →

Craft & detail Critique & decision-making Prototyping & exploration

NIELSEN NORMAN GROUP

LDLR DESIGN FIELD NOTES

A demo is not a distribution

Budiu translates eval rigor into UX-research grammar — representative inputs, repeated runs, confidence intervals — and retires the one-good-output proof for good.

BY RALUCA BUDIU / AUG 14, 2026

One AI Output Is an Example, Not an Evaluation

A single good output is like a participant who completes a task: encouraging, but not an evaluation.

Why it matters

— Ten representative inputs, five runs each: fifty answers before judging

— Two systems can share an 80% score and fail in opposite ways

— Test-input and run-to-run variability answer different questions; neither substitutes

— For nondeterministic systems, a passing QA test is a sample, not proof

— If runs are too expensive to repeat, the evidence is insufficient

Read on Nielsen Norman Group →

Research & validation Critique & decision-making

DESIGN SYSTEMS COLLECTIVE

LDLR DESIGN FIELD NOTES

Local first, system later

A case for local components as signals, not violations — designers build in place, and the system team governs what earns promotion to the library.

BY VEDANT / AUG 21, 2026

Let Designers Build. Let Design Systems Govern.

The maturity of a Design System isn't measured by how few local components exist.

Why it matters

— Local components are experiments; system components are governed contracts

— Six steps: build, validate in product, review, assess, contribute, publish

— Five teams building the same empty state is feedback, not violation

— Custodian becomes governor and enabler; designer becomes contributor

— Not every successful local pattern deserves promotion to the system

Read on Design Systems Collective →

Systems & primitives Cross-functional & process Critique & decision-making

NIELSEN NORMAN GROUP

LDLR DESIGN FIELD NOTES

Own the string, not the wind

Sunwall names the attachment that burns designers out — ownership of things they can't control — and shows where to reattach it, scenario by scenario.

BY EVAN SUNWALL / AUG 14, 2026

Psychological Ownership: Own the Right Things

You can't switch the feeling off, but you can influence what it holds on to.

Why it matters

— Ownership grows from control, deep knowledge, and invested effort

— Relief at working alone is ownership attaching to authorship, not outcomes

— Swap 'my design' for 'our design' — language redirects the attachment

— The build diverged? Document UX debt before assigning character flaws

— Fully yours: judgment, actions, growth, values, boundaries, relationships

Read on Nielsen Norman Group →

Hiring & team Critique & decision-making

LINEAR

LDLR DESIGN FIELD NOTES

Fence the field, don't script the path

Wolting explains how Linear replaced predictability with boundaries — system prompt, tool design, run scope, custom harness — and let the agent find its own way inside them.

BY MATTHIJS WOLTING / AUG 10, 2026

How we built Linear Agent

So rather than engineering a fixed path for Linear Agent, we defined the boundaries within which it could find its own.

Why it matters

— Five boundary surfaces: system prompt, tools, product model, run scope, custom harness

— Constraints live in tool design, not prompt prose — the tool can't overreach

— Coding agents get few deep tools; Linear Agent gets many shallow ones

— System skills disclose prompt fragments and tools progressively, preloaded by inference

— They refused the agent the SDK and GraphQL API — breadth traded for blast radius

Read on Linear →

Systems & primitives Critique & decision-making

DESIGN SYSTEMS COLLECTIVE

LDLR DESIGN FIELD NOTES

Three truths, no liar, one referee

Williams calls the end of the single source of truth: Figma holds intent, Storybook behavior, production reality — and reconciling them is a job nobody has posted yet.

AUG 5, 2026

The Job Nobody Is Hiring For Yet

Evidence gets faster and cheaper. The verdict still needs someone who can be asked why, in public, and answer.

Why it matters

— The opening incident: Figma showed four dropdown actions, Storybook four, production two

— None of the three lied — engineering removed actions that generated support tickets

— AI makes cross-checking the layers near-free; deciding which layer governs stays human

— Smarter or rotting? Telling those apart is judgment, and judgment needs an owner

— Most real design-system teams run three to five people, understaffed by their own report

Read on Design Systems Collective →

Systems & primitives Critique & decision-making Hiring & team

UX COLLECTIVE

LDLR DESIGN FIELD NOTES

Rotate the point of the V

Yeo turns murmurations and goose formations into an operating model — async telemetry, rotating leads, domain pods, three anchors, and a scheduled molt for design debt.

AUG 13, 2026

Rethinking design leadership with swarms and flocks

Nobody can break wind at the front indefinitely without burning out.

Why it matters

— Starlings track only six or seven neighbors; V-formations save 14% metabolic energy

— Honking comes from the back: swap status meetings for async telemetry, 24-hour replies

— Led two launches back-to-back? Hand off the point, and protect it in calibration

— Pods of three or four designers per domain, with a Staff-level craft guardian

— Cap non-negotiable principles at three; schedule a flightless molt for design debt

Read on UX Collective →

Hiring & team Cross-functional & process Critique & decision-making

SLACK DESIGN

LDLR DESIGN FIELD NOTES

Adequate became free; loved didn't

Fernandez turns Butterfield's 2013 promise into a working framework — utility, usability, feel — with named failure modes and a gut-check: are we proud to ship this?

BY MIGUEL FERNANDEZ / JUL 17, 2026

Product Quality: A Shared Commitment to Craft in the wake of AI

The distance from nothing to adequate has collapsed. The distance from adequate to loveable has not.

Why it matters

— Craft debt compounds like technical debt; you pay in churn and reputation

— Utility, usability, feel: one craft language across product, engineering, design

— The IKEA effect at scale: we overvalue whatever we prompted

— Looks-finished is not is-finished; visible behavior hides absent structure

— The bar: would I recommend this to someone I care about?

Read on Slack Design →

Craft & detail Critique & decision-making Cross-functional & process

DESIGN SYSTEMS COLLECTIVE

LDLR DESIGN FIELD NOTES

The rulebook compiles now

Williams, from Morales's talk: latency was your real reviewer. When agents erase the wait, the checking must be encoded — governance shipped as an output, not a fight.

JUL 13, 2026

Governance Becomes an Output

Governance doesn't have to be a manual fight. It just has to be encoded.

Why it matters

— Slow production looked like discipline; mostly it was latency doing review

— Remove the clock and the checking goes undone, loudly, all at once

— Morales's build: encode conventions into agents so governance ships as output

— Infrastructure can only encode what the team actually agreed on first

— Ask first: does your system know what it is?

Read on Design Systems Collective →

Systems & primitives Critique & decision-making

GITHUB

LDLR DESIGN FIELD NOTES

The first patch is a price check

The forty-minute scope debate now costs more than generating the diff. Abuadas's move: produce a constrained probe, then argue from evidence instead of vibes.

BY DALIA ABUADAS / JUL 17, 2026

The cost of saying yes has changed

A thousand-line diff that technically passes but nobody wants to own is not a cheap change. It's a deferred cost.

Why it matters

— The most expensive part of a small request is now the meeting about it

— Generate the first patch as a probe; argue from the diff, not from vibes

— Four-line diff with passing tests: ship it; touches auth middleware: it was never small

— Constrained attempts: smallest patch, behind the flag, contract unchanged, every touched file listed

— The new skill is pricing uncertainty fast, not defaulting to yes or no

Read on GitHub →

Critique & decision-making Cross-functional & process

NIELSEN NORMAN GROUP

LDLR DESIGN FIELD NOTES

Product sense is a closed loop

Kohler rebuts the empathy-plus-creativity definition: sense is pattern recognition built by facing problems, shipping solutions, and staying to see the results.

JUL 24, 2026

A Concrete Definition of "Product Sense" (and How to Build It)

AI may rob you of the chance to develop product sense if you aren't careful.

Why it matters

— Against Walter's empathy-plus-creativity: sense is pattern recognition from completed loops

— Fire chiefs, NICU nurses, chess grandmasters: intuition built by seeing outcomes

— Wicked situations feel like pattern matches and reinforce the wrong lesson

— Stay long enough to see results; document your logic before looking at them

— Half-loop builders feel experienced but never learn whether the building collapsed

Read on Nielsen Norman Group →

Critique & decision-making Research & validation

PROOF OF CONCEPT

LDLR DESIGN FIELD NOTES

Step into the pocket, don't scramble

Quarterbacks under a blitz buy time by stepping toward pressure, not fleeing it. Hoang maps the move onto tech's three blitzes and the drift that follows panic.

JUL 26, 2026

Climbing the pocket of pressure

Scrambling out of the pocket creates more vulnerabilities for the quarterback and often creates chaos from the originally intended play.

Why it matters

— Quarterbacks buy time by stepping toward pressure, not fleeing it

— Three blitzes named: competitors on your roadmap, more-with-less, shifting priorities

— Internal pressure comes up the middle; speed collapses the pocket faster

— Avoid the drift: movement without progress, staring at the rush not the field

— Defer blame until the play ends; retrospectives don't steer sinking rafts

Read on Proof of Concept →

Critique & decision-making Hiring & team

PROOF OF CONCEPT

LDLR DESIGN FIELD NOTES

Taste is a training regimen

Hoang rejects taste-as-gift and lays out four deliberate practice areas — eye, hand, aesthetics, rhetoric — for training design judgment like a discipline.

JUL 19, 2026

Training design senses

Experience isn't time spent; it's evidence of what you can do.

Why it matters

— Agentic tools without foundations: a DJ set handed to a pianist-in-training

— Train the eye by capturing, collecting, and curating what you notice daily

— Copy the masters: reconstruct interfaces from scratch and ask why each margin works

— Aesthetics come third, once the eye and the hand can keep pace with each other

— Rhetoric sharpens judgment: if you can't explain a call, intuition is ahead of reasoning

Read on Proof of Concept →

Craft & detail Critique & decision-making Storytelling & narrative

MC DEAN

LDLR DESIGN FIELD NOTES

Abundance is the new management problem

Production capacity stopped being the constraint. Dean maps what design leaders manage instead: judgment, team cohesion, and the moment to stop generating.

JUL 23, 2026

Design Leadership Just Got Harder

Deciding what good means when the machines will hand you a thousand versions of fine.

Why it matters

— Anchored in the AI in Design Report: 900+ designers, leaders at Anthropic, Stripe, Shopify

— Designers reporting decreased collaboration jumped roughly four-fold in a single year

— Encode your taste so the team and the AI have something concrete to aim at

— Be the one who says stop; endless generation drowns teams in options

— Evaluation systems still measure the old job while expecting AI-native judgment

Read on Substack →

Hiring & team Critique & decision-making Cross-functional & process

DESIGN SYSTEMS COLLECTIVE

LDLR DESIGN FIELD NOTES

Execution got cheap. Understanding didn't.

A designer watches AI produce complete outcomes without touching the reason they exist, and argues design's scarce skill is now interpretation.

JUN 22, 2026

AI Didn't Break Design. It Exposed It.

Execution is becoming cheap and abundant. The real scarcity is understanding.

Why it matters

— AI slipped into design's own layer: pattern recognition and structured decision-making

— Treating AI as a tool locks attention on prompts; it is a meaning-making layer

— Context can be engineered; intent stays unspoken and is harder to construct

— Design systems shift from systems of appearance toward systems of understanding

— The value question isn't which tool next, it's where your value sits now

Read on Design Systems Collective →

Systems & primitives Critique & decision-making

GEOFFREY LITT

LDLR DESIGN FIELD NOTES

Understand to participate

Litt argues understanding agent-written code keeps you fluent enough to have the next idea, and offers four techniques for getting there.

JUL 2, 2026

Understanding is the new bottleneck

A quiz is a speed regulator.

Why it matters

— Agents now verify their own work, so understanding buys creative participation instead

— A project is many loops, and fluency is what lets you have the next idea

— His /explain-diff skill outputs literate diffs: background first, intuition before code

— Every explainer ends in a five-question quiz he won't ship without passing

— Micro-worlds: he had an agent build a debugger so he could learn the system himself

Read on their blog →

Critique & decision-making Prototyping & exploration Cross-functional & process

UX COLLECTIVE

LDLR DESIGN FIELD NOTES

The what slips, one decision at a time

A product designer with no technical background shipped an agent that grades agents, and felt the what start sliding toward the machine.

BY FRANCISCO BARRERA AROS / JUL 10, 2026

The crisis of the what

Nobody handed the what to AI in a strategy meeting.

Why it matters

— He's a product designer with no technical background who shipped an agent that grades agents

— The how was always negotiable; he argues the what is now in play too

— He parts ways with Thompson's map, which holds the what until superintelligence

— The handover comes through small decisions that look like pure execution

— What survives, in his read: judgment built through experience, intuition with scars

Read on UX Collective →

Critique & decision-making Prototyping & exploration

FIGMA

LDLR DESIGN FIELD NOTES

Seven questions design leaders are asking now

Figma distills two days with 1,300 senior leaders into the questions that matter: expertise, team shape, trust, and the quality bar.

BY EMMA WEBSTER

7 questions we had going into Config Leadership Collective

"Expertise is no longer having the answer, because there will always be more questions." (Sneha Narahalli, Sephora)

Why it matters

— Expertise is shifting from execution toward judgment, taste, and discernment

— Airbnb's Teo Connor rebuilt her org into startup-style pods with editing eyes

— Google's Sheta Chatterjee: fundamentals haven't changed, users and craft still lead

— Agent trust gets built gradually, with approval steps that relax over time

— Cisco's Jeetu Patel on quality: taste can't be handed to an agent

Read on Figma →

Critique & decision-making Hiring & team

FIGMA

LDLR DESIGN FIELD NOTES

Steady is a leadership strategy

Jen Dunnam on leading through acceleration: less panic, harder interview probes for dissent, and naming which speed actually matters.

BY EMMA WEBSTER

How to keep design teams steady through constant change

"The mistake is thinking speed is one thing." (Jen Dunnam, Hinge)

Why it matters

— Her job is steadying teams, not manufacturing urgency they already have

— Interviews probe for dissent: where did you push back, and why

— A team nodding along signals disengagement, confusion, or fear

— Crit is where leaders model questioning their own decisions

— Fast, cheap, good: the real work is naming which two

Read on Figma →

Hiring & team Critique & decision-making

LINEAR

LDLR DESIGN FIELD NOTES

Review Is Judgment Now, Not Parsing

A Linear engineer on how agent-heavy review freed him from parsing lines to judging whether the work earns its place.

BY MACIEK PEKALA / JUN 10, 2026

Reviewing code in the agent era

this shift toward judgment is the one I actually see in my own work.

Why it matters

— PRs merged jumped 50% in a quarter, and rubber-stamping started to tempt him.

— Agents now cover correctness, so his reclaimed time goes to system fit.

— He asks whether the change solves the right problem, not just whether it works.

— Review becomes the rare moment teammates look at each other's work directly.

— Quality shifts from correct to useful: does this earn its place?

Read on Linear →

Critique & decision-making Cross-functional & process

DESIGN SYSTEMS COLLECTIVE

LDLR DESIGN FIELD NOTES

Encode Intent, Not Just Components

When agents author your design system, intent must live in tokens and metadata, not only in a practitioner's head.

JUN 29, 2026

The System Knows What You Built. Does It Know What You Meant?

AI has lowered the floor on execution. The ceiling on judgment is still yours.

Why it matters

— The question moved from can AI use your system to can it write it.

— Agents connected to metadata make decisions about what to reach for.

— Discoverability isn't documentation coverage; it's intent legibility where decisions get made.

— Tokens should carry purpose, not just values; metadata should explain what's for.

— Unresolved ambiguity gets filled by the agent's best guess, everywhere, unnoticed.

Read on Substack →

Systems & primitives Critique & decision-making

UX.RASPBERRY

LDLR DESIGN FIELD NOTES

From producing to deciding

A blunt read on where designer value sits in 2026: production is getting automated, so strategy and judgment are what stay human.

MAY 1, 2026

UX Design 2026+: From producing to deciding, from craft to operation

It's no longer about what you produce; it's about what you decide.

Why it matters

— The tedious middle of the job is being automated away.

— Value moves to the before (strategy) and the after (judgment).

— The "Junior Paradox": one senior plus AI replaces two juniors.

— Survival looks T-shaped, with one deep, hard-to-automate strength.

Read on Medium →

Critique & decision-making Hiring & team

DESIGN MAVERICKS

LDLR DESIGN FIELD NOTES

Design is becoming leadership literacy

Forty design leaders in Amsterdam reach the same conclusion: the role is dissolving while the capacity becomes foundational to leadership.

JUN 4, 2026

When AI Builds Everything, Design Becomes Everything

We are training people for a job. We should be cultivating a capacity.

Why it matters

— Intercom now ships over 90% of pull requests authored by Claude Code

— When execution costs nothing, judgment about direction becomes the product

— Atlassian runs 64% ICs, every designer a platform designer, demo before memo

— Frigieri's upstream conditions: workflows, data pathways, decision rights, controls

— The Frozen Middle: everyone experimenting alone without strategic guidance from above

Read on Design Mavericks →

Hiring & team Critique & decision-making Systems & primitives

UX COLLECTIVE

LDLR DESIGN FIELD NOTES

A skeptic's playbook for AI

A research leader of eight argues for measured AI adoption: pressure-tested tools, tracked outputs, and guardrails before enthusiasm.

APR 30, 2026

Thoughtful AI implementation for UXR leaders

I'm arguing for a measured approach to their usage.

Why it matters

— Pressure-test NotebookLM with likely stakeholder questions before opening access

— Capture before and after results when tuning prompts with ops

— Track who queries research repositories and how outputs shape decisions

— Half of Americans are more concerned than excited about AI

Read on UX Collective →

Research & validation Critique & decision-making

THE MERHOLZ AGENDA

LDLR DESIGN FIELD NOTES

The Layoffs Are Social Contagion

Merholz reads the layoff wave through decades-old org theory: under uncertainty, executives copy each other, then call it AI strategy.

MAY 23, 2026

Monkey see, monkey mandate

The uninformed copy the foolhardy.

Why it matters

— Pfeffer named it years ago: layoffs spread by imitation, not evidence.

— Mimetic isomorphism, copying peers under uncertainty, explains RTO, AI mandates, and cuts.

— Token-spend quotas reward activity over impact, like counting lines of code.

— CEOs trust peers in other contexts more than the teams who know their own.

— The cost lands on people who were never in the room.

Read on The Merholz Agenda →

Critique & decision-making Cross-functional & process Hiring & team

POSTHOG

LDLR DESIGN FIELD NOTES

When FOMO Ships It Too Early

PostHog's CEO on shipping AI before it worked, the architecture fix that turned it around, and betting the product on agents.

BY JAMES HAWKINS / MAY 1, 2026

PostHog's next chapter

FOMO came ahead of progress.

Why it matters

— A one-shot assistant shipped early, returned empty graphs, and felt half-useful.

— The fix was a single query loop so the agent could iterate.

— Agents now create most new dashboards, with MCP usage roughly doubling monthly.

— The bet: a headless, proactive product that ships fixes while you sleep.

— Hawkins is candid about UX churn, and why the short-term pain is worth it.

Read on PostHog →

Prototyping & exploration Critique & decision-making Systems & primitives

UX COLLECTIVE

LDLR DESIGN FIELD NOTES

When the Metric Becomes the Harm

Bertini argues engagement-maximizing AI doesn't fail by accident, and that designers and product leaders own the direction it takes.

MAY 4, 2026

We built this. Now we own it.

That is not a flaw in the design. It may be the design.

Why it matters

— Klarna's bot optimized for throughput over empathy, and the gap shows.

— A companion chatbot getting it wrong carries far higher stakes than a refund delay.

— Maximizing engagement leans into vulnerability, which may be intended, not accidental.

— Harari's warning: autonomous systems may not give us a second chance.

— Responsibility for AI's direction sits with the people building it.

Read on UX Collective →

Critique & decision-making Storytelling & narrative

DESIGN SYSTEMS COLLECTIVE

LDLR DESIGN FIELD NOTES

How to Position System-Level Design Work So It Actually Gets Funded

Frame it as infrastructure, not UX, or it gets deprioritized before the meeting ends. · If you don't define the cost of inaction, leadership assumes there isn't one.

MAY 15, 2026

How to Position System-Level Design Work So It Actually Gets Funded

Why it matters

— Frame it as infrastructure, not UX, or it gets deprioritized before the meeting ends.

— If you don't define the cost of inaction, leadership assumes there isn't one.

— The ask isn't an improvement; it's a block on compounding operational inefficiency.

— Four things need to be undeniable: the problem, the cost, the scope, and governance.

— Portfolio-level framing gets director attention; product-level framing gets project-scoped budget.

Read on Design Systems Collective →

Cross-functional & process Storytelling & narrative Critique & decision-making

BOOTCAMP (MEDIUM)

LDLR DESIGN FIELD NOTES

The Design System You Didn't Choose

shadcn became the default behind every AI-generated UI. A developer built it. Most design teams never noticed they're now competing with it.

MAY 11, 2026

The most important Design System in 2026 that designers missed was built by a developer.

Every shadcn project is a fork.

Why it matters

— AI tools converged on shadcn as the default shape of generated UI.

— It lives in GitHub and React templates, not Figma, so designers missed it.

— The MCP server turned accidental convergence into installed infrastructure.

— The work isn't fighting the default. It's reading it, then deciding what to override.

Read on Bootcamp (Medium) →

Systems & primitives Critique & decision-making

GITHUB (ENGINEERING & DESIGN)

LDLR DESIGN FIELD NOTES

When 'feels instant' is the spec

Most performance work chases tail latency. GitHub Issues reframed it around distribution shape, then named 4.7% server/cache divergence as the price of feeling instant.

BY ALEXANDER LELIDIS / MAY 14, 2026

From latency to instant: Modernizing GitHub Issues navigation performance

We are willing to serve data that may be slightly stale if that allows the navigation to complete near instantaneous

Why it matters

— Reframed the goal: shape the distribution, not chase p99.

— Five layers: in-memory cache, IndexedDB, preheat, service worker, code-split.

— Accepted 4.7% server/cache divergence as the operating envelope.

— Cache hit went from 33% to 96%. P10 dropped to 70ms.

Read on GitHub (Engineering & Design) →

Craft & detail Systems & primitives Critique & decision-making

DESIGN SYSTEMS COLLECTIVE

LDLR DESIGN FIELD NOTES

AI didn't add to the spec

AI-readiness isn't a separate category. The same fundamentals (tokens, layout intent, accessibility, governance, handoff contracts) finally have to read clearly to a non-human.

APR 27, 2026

Legible by Design: What AI Readiness Actually Demands of a Design System

The best systems are the ones that remain coherent even when the people who built them have moved on

Why it matters

— AI doesn't need a new system. It exposes the broken one.

— Five load-bearing properties: tokens, layout intent, a11y, governance, handoff.

— Legibility to machines reveals what was always opaque between humans.

— Stewardship beats generation. The system has to outlive its authors.

Read on Design Systems Collective →

Systems & primitives Critique & decision-making Cross-functional & process

WEBFLOW

LDLR DESIGN FIELD NOTES

Where the AI code debt lands

AI-generated code defers cost, it doesn't reduce it. Seven debt categories concentrate far from where the velocity gain was booked, including reviewer fatigue and security debt.

BY ANKIT AGRAWAL / MAY 12, 2026

The cleanup cost of ungoverned AI-generated code

It costs a contributor five minutes to generate a low-quality AI pull request, and hours for the maintainer to verify and reject it.

Why it matters

— Seven debt categories, segmented by who actually pays the bill.

— Time-to-exploit window collapsed from months to days.

— Five minutes to generate a PR. Hours to reject one.

— Triage PRs by risk class. Output is the wrong metric.

Read on Webflow →

Critique & decision-making Cross-functional & process Systems & primitives

MICROSOFT DESIGN

LDLR DESIGN FIELD NOTES

When metrics optimize the wrong thing

Frontier AI teams optimize what's measurable, not what matters. Microsoft's user researchers want a seat at the evaluation table, as a safeguard against the metric chosen for being convenient.

BY MARTIN LINDE DANTY, RABEEZA A., ANNE KLÆRKE-OLESEN, MADHUMAY SINHA, MICHAEL O'SULLIVAN / APR 28, 2026

The missing seat at the frontier team table

When evaluation is driven primarily by available data, teams risk optimizing what is measurable rather than what actually matters.

Why it matters

— User research belongs in evaluation, not adjacent to it.

— Session-length looked strong. Task-completion cut repeat queries by a third.

— Sales-order workflow assumed isolation. Research surfaced cascading approvals.

— Concept tests caught a workflow misread that would have cost weeks.

Read on Microsoft Design →

Research & validation Cross-functional & process Critique & decision-making

LINEAR

LDLR DESIGN FIELD NOTES

Linear's hiring playbook, written down

Most companies don't write down how they hire because writing it down forces honesty about what they actually value. Linear did the work and published it.

BY DOUG PARKER / APR 28, 2026

Hiring & team Cross-functional & process Critique & decision-making

LINEAR

LDLR DESIGN FIELD NOTES

Coding agents kept in their lane

An engineer at Linear scopes coding-agent tasks tight enough that they can't wander. The discipline is the same as scoping for a junior teammate.

How we use Linear Agent at Linear

Matthijs gets the best results from coding agents by breaking the work into small, targeted steps.

Why it matters

— Scoping for an agent is the same discipline as scoping for a junior.

— Three-team workflow shows where agents add real leverage.

— PM-ships-the-fix works because the change was small, not the agent.

— Marketing-shaped piece. The underlying tactics are still real.

Read on Linear →

Cross-functional & process Critique & decision-making

MAGGIE APPLETON

LDLR DESIGN FIELD NOTES

When coding stops being the bottleneck

Maggie Appleton reads Steve Yegge's chaotic agent orchestrator as speculative design fiction. The coding bottleneck is gone; the design bottleneck was always there.

JAN 23, 2026

Gas Town’s Agent Patterns, Design Bottlenecks, and Vibecoding at Scale

Design becomes the limiting factor: imagining what you want and figuring out the gnarly details to make it real.

Why it matters

— When agents write faster than you design, vibe-designing is a footgun.

— Mayor/Polecat hierarchy sketches a real coordination pattern for agents.

— Code distance shifts with domain, stakes, and team size.

— Infrastructure matters more than models. Validation loops, not smarter agents.

Read on their blog →

Critique & decision-making Cross-functional & process Prototyping & exploration

MAGGIE APPLETON

LDLR DESIGN FIELD NOTES

The PR is not a planning tool

Coding agents collapsed the implementation window. The alignment checkpoints that used to live inside it just disappeared, and the PR cannot carry their weight.

APR 13, 2026

One Developer, Two Dozen Agents, Zero Alignment

Code is so cheap we don't properly stop to think before prompting it. We lose more alignment points.

Why it matters

— Planning and building are no longer separate phases. The tools still are.

— The PR now carries upstream checkpoints it was never designed for.

— Most alignment context lives in people's heads. Agents can't reach it.

— Ace is a research prototype. The diagnosis is sharper than the fix.

Read on their blog →

Cross-functional & process Critique & decision-making Prototyping & exploration

RAUNO FREIBERG

LDLR DESIGN FIELD NOTES

The minimap that became a blur

Rauno tried to solve 'scroll back' with a minimap. The minimap caused jank, the jank needed a blur, and the blur almost worked.

Field Notes

The jank gets increasingly worse when you actually choose to scroll to another message.

Why it matters

— Each solution reveals the next constraint: scroll, minimap, jank, blur.

— AI summaries failed: you remember what you typed, not a summary.

— Blur-and-stack beats layout collapse because fewer objects move.

— Keeping the composer visible preserves task context in minimap mode.

Read on their blog →

Critique & decision-making Prototyping & exploration Craft & detail

LINEAR

LDLR DESIGN FIELD NOTES

Coding agents kept in their lane

An engineer at Linear scopes coding-agent tasks tight enough that they can't wander. The discipline is the same as scoping for a junior teammate.

BY RHEA PUROHIT / APR 10, 2026

How we use Linear Agent at Linear

Matthijs gets the best results from coding agents by breaking the work into small, targeted steps.

Why it matters

— Scoping for an agent is the same discipline as scoping for a junior.

— Three-team workflow shows where agents add real leverage.

— PM-ships-the-fix works because the change was small, not the agent.

— Marketing-shaped piece. The underlying tactics are still real.

Read on Linear →

Cross-functional & process Critique & decision-making

NUBANK DESIGN

LDLR DESIGN FIELD NOTES

Killing darlings, statistically

Nubank ran a real A/B test on whether to scale or shut down an internal HR product. The data was unambiguous; the team's reaction to it was not.

BY NUBANK EDITORIAL / AUG 27, 2025

A/B Testing for Decision of scaling or decommissioning an Human Resources product

Increased usage didn't always translate to feeling recognized — a critical insight missed without experimental

Why it matters

— A/B testing applied to decommissioning, not just feature launches.

— Internal tools deserve the same rigor as customer-facing products.

— Usage below threshold: kill it. Above: scale it.

— Hardest part isn't the data, it's letting the team accept it.

Read on Nubank Design →

Research & validation Cross-functional & process Critique & decision-making