Working Surface

1 October 2026: 3 links

2026-10-01T00:00:00Z

Three pieces from the past four days place the risk of AI-assisted work at the moment a person must explain or judge what an AI agent produced. An engineering leader, a usability researcher and a summit of product leaders each argue that understanding and judgement do not grow with output and have to be checked on purpose.

LeadDev: Your error budgets don't know AI exists. Paul LaPosta, a DevOps leader writing in LeadDev, an engineering leadership publication, argues that AI makes changes cheap to produce but not to understand.
UX Tigers: User Vigilance Fails Twice in the AI Age: Watch Duty & Verdict Duty. Jakob Nielsen, a usability researcher, argues that people approving an AI agent's actions stop paying attention, and that products should ask less often and offer undo.
Lenny's Newsletter: All of the Lenny & Friends Summit talks are now online!. Lenny Rachitsky, who writes Lenny's Newsletter for product managers, reports that summit speakers disagreed on process but agreed that judgement matters more as AI makes building easier.

30 September 2026: 3 links

2026-09-30T00:00:00Z

Dropbox, Shopify and Figma describe what changes once AI agents make building cheap. Dropbox checks agent changes like a person's, Shopify rebuilds its apps natively with one shared test suite, and Figma's product lead says choosing a direction is now the hard part.

Dropbox.Tech: Evolving our calendar assistant Reclaim to be AI-native without starting over. Dropbox's engineers sent every AI agent change in Reclaim, their calendar assistant, through the same checks of scheduling rules that people's changes pass, and let users preview it first.
The Pragmatic Engineer: Why has Shopify dropped React Native?. Shopify is rebuilding its React Native mobile apps as separate iOS and Android apps, because AI coding agents have made building each feature twice cheap enough.
Figma Blog: What it takes to build great products now. Paige Costello, vice president of product at Figma, argues that AI makes prototypes quick, so the hard work moves to choosing a direction, finishing the product and making it distinctive.

29 September 2026: 11 links

2026-09-29T00:00:00Z

When AI tools do more of the drafting and building, some work that people did along the way stops happening. Six pieces name parts of it, from practising a skill to writing down why a rule exists, and suggest how a team can keep doing that work. Five engineering pieces from 29 September cover teams of coding agents, architecture, a developer survey, earlier approvals and how junior engineers build judgement.

this+that: Nobody notices when a company forgets. Jeff Reynar, chief executive of the AI start-up this+that, argues that when AI agents do the work, people's skills and a company's knowledge can fade without anyone noticing.
The Product Picnic: LLMs are just normal technology. But tech is just a normal medium.. Product writer Pavel Samsonov argues that AI prototypes tempt teams to skip the first step in making a product: deciding what problem it should solve.
Nimble UX: Where do we find room to think?. Maya Elise Joseph-Goteiner, founder of the user-experience agency Velocity Ave, worries that research teams increasingly edit AI output, which cannot push back as a colleague would.
Design Systems Collective: Before You Let an AI Agent Touch Your Design System, Write Down Four Rules. Guilherme Negreiros, who builds a design system with AI agents, writes every exception into the short rule files the agents read, because an automated audit removed a deliberate browser fix.
Nate's Newsletter: Executive Briefing: You Bought Better Tools and Your Finished Work Still Waits. AI newsletter writer Nate B. Jones argues that one person's speed with AI agents can outrun a team's review and decisions, and names six principles for raising team output.
Design Systems Collective: “You don’t fix the output, you fix the system” — Cristian Morales Achiardi on agentic design systems. Cristian Morales Achiardi, a design engineer, generates a design system's code, design files and documentation from specifications machines can check, so a wrong result is fixed in those specifications.
Medium: Durable Agentic Teams: Maximizing Your Software Development Workflow. Jonah Turnquist, chief technology officer of the property-planning software company PropCode, organises his AI coding agents like an engineering team, and finds that his own review time becomes the limit.
Accidentally in Code: Good Architecture Is a Function of Time. Cate Huston, who led the rebuild of Twill, a recruiting platform, says a design system and a strict permissions model paid off once her team built with AI agents.
Agoda Engineering & Design: Key Takeaways from Agoda’s AI Developer Report 2026. Agoda, the online travel company, surveyed developers in Southeast Asia and India: 53 percent use AI agents widely, and 79 percent require a person to approve production deployments.
Medium: The Approve Button Is Not a Control. Rama Lingamgunta, who builds AI agent platforms, replaced a single Approve button with three earlier approvals after a reviewer passed a 1,400-line agent change in under two minutes.
LeadDev: AI makes critical thinking harder to build. Mia de Búrca, a staff engineer at Vistaprint, argues that AI tools spare junior engineers the struggles that built their judgement, so leaders should make room for that practice.

28 September 2026: 10 links

2026-09-28T00:00:00Z

Five pieces from designers and engineers show which jobs people keep when AI agents do the building. They keep the design system and the first screen, the review of what agents make, the files agents read, and responsibility for the product as a whole. Five engineering pieces from 28 September cover reviewing AI-written code, agents that change live systems, agents that build agents, new AI models and tests written first.

Behind the Craft: We Built Grok Bot. Here Are Our 14 Best Bots | Peng Zheng & Lauren Tan. Two leads at SpaceXAI show how they share work with AI agents: the designer builds the first screen by hand, and the engineer's agent divides projects among other agents.
Design Outcomes: New Patterns Are a Tax. A design team for clinical software drafted two rules for keeping its product simple, reuse existing patterns and judge each design in context, from reviews that included AI-generated layouts.
Design Outcomes: Good Isn't Great. A designer fixed a billing problem for clinicians in an afternoon with an AI coding agent, and his design lead used the review to take it from good to great.
Atlassian: Introducing CAFE(S): A framework for defining AI context quality. Brian Houck of the developer-research firm DX defines five qualities of the instructions AI agents read, and argues that each shared instruction file needs a team that owns it.
Sublime Coding: DHH's Rails World 2026 Keynote: Pencils Down, Now What. Engineering leader Jared Smith argues that when AI agents write the code, a person must still own how the changes fit together, which a better model will not fix.
Elevate: The Code Nobody Reads. Addy Osmani withdraws his advice to read every line of AI-written code and proposes automated review of every change, human review matched to risk, and a person approving every merge.
LeadDev: Your agent loop is not a production system. An Amazon Web Services engineer argues that once an AI agent can change live systems, teams must prove separately who approved a change, what ran and whether it worked.
Salesforce Engineering: Engineering Multi-Agent AI Teams That Build and Test Themselves. A Salesforce team built AI agents that design and test teams of other agents, aiming to cut hours to minutes, with an engineer approving every design.
Databricks: How Databricks rolls out frontier models to 12,000 employees on Day 1. Databricks, a data and AI software company, gives staff each new AI model on release day within a capped budget, then keeps or drops it on measured cost and quality.
BLOG@CACM: Nobody Did TDD for 25 Years. Now the Machine Requires It. Abtin Aghagolian, chief technology officer of Pikd, a London augmented-reality company, argues that tests written before an AI agent writes code are now how engineers control its output.

27 September 2026: 5 links

2026-09-27T00:00:00Z

When AI writes part of a team's work, reviewers and users need to see what was made, by whom and from what. The pieces label sections by author, link summaries to sources, keep one document per feature, show users when a save happens, and ask for expert reviewers.

Design Outcomes: Know What to Doubt. A design team labels each part of a ticket for an AI coding agent as human-written or agent-written, so that reviewers know which kind of error to look for.
The Beautiful Mess: TBM 441: AI, the Loss of Positive Friction, and What to Do About It. Product writer John Cutler argues that AI removes the manual steps where teams used to judge customer feedback, and proposes keeping every summary linked to what customers said.
Design Outcomes: Three Lattes, One Order. A design lead proposes keeping one requirements document per feature, a table with a prototype under each row, so that people and AI agents build from the same version.
SVPG: Experts Lead Experts. Marty Cagan of the product consultancy SVPG argues that the current push for fewer managers is temporary, and that each craft still needs leaders who are expert in it.
Design Outcomes: A Working Feature Isn't the Same as a Trusted One. Clinicians doubted an autosave feature that worked, so the product team paused its rollout and added an indicator that shows the save while it happens.

26 September 2026: 7 links

2026-09-26T00:00:00Z

Six pieces describe how teams change their work once AI agents can finish whole tasks. They show the result before the tool, take workflows apart before automating them, set checks before release and make people responsible for managing agents. One engineering piece from the same day asks whether engineers still enjoy the work once their team codes with AI.

LukeW: Skip the Tools, Make the Outcomes. Designer Luke Wroblewski found that readers of his AI newsroom mostly asked for a report on their question instead of browsing articles, and argues for outcomes before tools.
GitHub Blog: When chat is the wrong UI. Burke Holland of GitHub argues that once a person knows a repeated job, having an AI agent build a small app for it costs fewer tokens than asking in chat.
Microsoft Inside Track: From the field: How agentic AI is reshaping adoption at Microsoft. Microsoft reports that staff in its Europe South region now take up AI agents without being pushed, and that the hard part has moved to governing them.
Microsoft Inside Track: Driving seller adoption of Sales Agent through role-based change management at Microsoft. Microsoft had a business programme manager take its sellers' workflows apart before rebuilding its sales AI agent, and is moving to judge the agent by deal speed rather than usage.
The Pragmatic Engineer: Design Engineering with Maggie Appleton. Maggie Appleton, a research engineer at GitHub, no longer reads the code AI agents write for her prototypes and writes a spec saying how the agent will check its work.
CIO: I stopped asking my team to use AI. I asked them to manage it. Dan Graves, chief product officer at WitnessAI, has seven people on his team manage AI agents like junior hires, with a person approving work before it moves on.
shubs.io: do we still enjoy software engineering in the age of AI?. Shubham Shah, co-founder of the company Assetnote, published an engineer's message about losing satisfaction since the team moved to coding with AI, and his reply as their manager.

25 September 2026: 5 links

2026-09-25T00:00:00Z

Three pieces look at what shapes AI work before anyone chooses: the defaults a product ships, the options a team rejects, and the material an AI can reach. Each argues that these inputs should be kept in view, and one argues that each needs a named owner. Two engineering pieces from the same day look at how AI tools change an engineering manager's job and how a large embedded-software company plans to reorganise around AI agents.

UX Tigers: Default Dominance: Most Users Never Change the Setting You Ship. Jakob Nielsen shows that most users keep whatever setting a product ships, and argues that AI agents hide their defaults in instruction files that need an owner.
Product Talk: Trash Can Tracking — All Things Product Podcast with Teresa Torres and Petra Wille. On the podcast All Things Product, Petra Wille proposes that teams keep a visible record of rejected problems and solutions; an empty record of rejected solutions means nobody compared options.
Nate's Newsletter: Your agent is giving you reasonable answers from half your context. A conversation with OpenAI.. Nate B. Jones reports from OpenAI that whether staff find AI useful depends more on whether it can reach their material than on their skill.
Codacy: How AI Is Changing the Engineering Manager Role: More Context, More Capacity, and the New Job of Protecting Focus. Two engineering managers at Codacy, a maker of code-quality tools, say AI tools let each engineer start several tasks at once, so they now limit how much work is open.
arXiv: Developing a Roadmap to an AI-first Organization: A Case Study in Embedded Software Development. Researchers held a workshop with 40 staff at a large embedded-software company and turned their views on AI agents into a roadmap for reorganising the company around them.

24 September 2026: 4 links

2026-09-24T00:00:00Z

Four pieces from Anthropic, Stripe, Cursor and Evil Martians describe setting the rules for AI agents' work before the work starts. The rules go into automatic checks, review policies, monitoring plans and the components themselves, and people judge what falls outside them.

Anthropic: How to prepare for AI-driven code modernization projects. Two Anthropic engineers advise companies modernising old code with AI agents to agree, before the work starts, what each change must prove and how much human review it gets.
Stripe: How Stripe is designing Checkout for AI agents. Stripe made its Checkout page cheaper for AI shopping agents to use by offering them only the actions valid at each step, built on the human checkout's own code.
Cursor: Bots for the last mile: Rollouts, Security Review. Cursor launched two AI bots for the work after a code change is proposed: one plans and watches the release, and one reviews every change for security flaws.
Evil Martians: AI makes design system guardrails mandatory; this framework delivers them. Evil Martians, a product consultancy, stops AI agents that build screens from changing the design system, and records every exception for a person to review.

23 September 2026: 4 links

2026-09-23T00:00:00Z

Four pieces argue that people can keep up with work made by AI agents only when it reaches them in small pieces that have already been checked. They also say that the judgment people bring has to be built first, by reading real cases and by working beside experienced colleagues.

Shopify Engineering: Helix: The internal tool powering our Shopify app's native migration. Shopify rebuilds its mobile app with AI models in small steps, each checked by tests, a visual comparison and two reviewing agents before an engineer approves it.
SEP: So, AI Killed Your Code Review Process. Now What?. An engineer at a software consultancy argues that when AI-written code outgrows review, teams should cut work into thin slices and review the agent's plan before any code exists.
Lenny's Newsletter: Advanced evals: How to find (and fix) hidden AI failures in your product. Two specialists in testing AI products argue that people should read real AI sessions and note the failures themselves before an agent looks for errors or anyone writes a metric.
UX Collective: In the age of AI, the UX field survives on leaders who cultivate juniors. A designer argues that design leaders should hire juniors again, because AI tools make newcomers useful sooner and questioning AI output is learned beside a senior.

22 September 2026: 4 links

2026-09-22T00:00:00Z

Four accounts from Warp, Linear and Bolt show how software teams check work now that AI agents produce it faster than people can review it. Each company moves part of the checking into written criteria, into files the agents read, or into a review split by expertise.

Lenny's Newsletter: How I AI: Meta's Muse review + How Warp ships 2,000 PRs a month with AI factories. Warp, which makes tools for software teams, lets the person who asked its AI agent for a code change review that change, instead of waiting for a separate reviewer.
Linear: AI coding has made CI a bottleneck, so we reworked ours to keep up. Linear rebuilt its automated code checks after AI agents made writing code faster than checking it, and taught its agents a new testing rule in the same change.
Warp: Using LLM-as-a-judge scoring to measure your software factory. Warp's chief executive explains how AI agents can grade a sample of coding agents' past work against written criteria, so that people review the failures rather than every change.
Bolt Labs: Design engineering at Bolt: shipping components with AI agents. At Bolt, designers now build interface components with AI agents directly in the library that ships, and engineers review those components instead of rebuilding them.

21 September 2026: 5 links

2026-09-21T00:00:00Z

Five pieces look at the work left for people once AI agents build quickly. It lies in reviewing agent-built changes, answering for what is handed over, and deciding what an interface promises its users.

Design Outcomes: The Bottleneck Moved to Review. When designers at Leonardo De La Rocha's company began making code changes with AI agents, the review broke down, so he set out who reviews them and to what standard.
The Product Picnic: Assigning agency to AI means surrendering your own. Pavel Samsonov argues that treating an AI model as a colleague lets people hand over work nobody has checked, and that whoever hands it over stays responsible for it.
GitHub Blog: Should you read the code, is RAG dead, and did Skills kill MCP?. A GitHub developer advocate argues that people must still review code written by AI, with effort matched to the risk, until they can explain and own the result.
Design Outcomes: What the Loading State Promises. A designer noticed that the loading text of his AI feature described several steps while the system made one quick call, and his design lead argued for a simpler indicator.
Design Outcomes: One Chat, and It Grows. Leonardo De La Rocha changed his mind twice in a week about how many AI chats his product should offer, and chose one because customers wanted one place to type.

20 September 2026: 2 links

2026-09-20T00:00:00Z

Two pieces look at the standing instructions people give AI agents before a task begins. Heavy users of one agent keep written rules, including which actions need their approval, and one engineer has his agent question him about each plan.

Nielsen Norman Group: The 3 Roles of Context for AI Agents. A Nielsen Norman Group study found that heavy users of the AI agent Claude sort its background information into three kinds, including standing rules on what needs their approval.
The Pragmatic Engineer: AI Skills with Matt Pocock. Engineer and educator Matt Pocock uses short instruction files that make AI coding agents question him closely about a plan, which leaves him more time for strategic work.

19 September 2026: 3 links

2026-09-19T00:00:00Z

Three pieces show that AI now produces working prototypes, whole apps and product comparisons in minutes. Each argues that the result is worth trusting only when something is settled first or can be checked: a written problem, fixed parts, or values with a source.

UX Tigers: The 100 Most Common UX Design Methods, Ranked by Value. Jakob Nielsen ranked 100 design methods by value and put AI prototyping first, while warning that a working demo argues as hard for a bad idea as a good one.
Product Talk: Creating Aha! Builder: Concept to Code, No Engineers Required. Aha!, a maker of product-planning software, built an AI app builder for product managers that creates a design system first and uses ready-made parts for sign-in and data.
UX Tigers: Comparison Tables Are Decision Machines: Design Them to Deliver a Verdict. Jakob Nielsen argues that comparison tables exist to help people decide, and that AI assistants now build their own, so tables need specific, sourced values in place of checkmarks.

18 September 2026: 3 links

2026-09-18T00:00:00Z

Three pieces show that a person's review of an AI agent's work is only as good as the reference that person checks against. Without a reliable reference, a reviewer can add errors instead of catching them.

DataHub: What Shipping an AI Agent to 16,000 People Taught Me About Reviewing One. Ananya Das, an intern at DataHub, tested a support agent before its release and found that an unchecked human review had made the agent look worse than it was.
DataHub: Context Engineering for AI Agents: Why the Hard Part Isn't the Context Window. DataHub, which sells data-management software, argues that AI agents give confident wrong answers because the company data they read was never checked, and that experts should approve it first.
The Airbnb Tech Blog: Beyond the model: Engineering AI infra with scientific judgement. Airbnb built Insight Miner, a tool that writes its analysis method into software around an AI agent, so that findings from customer-support data can be reproduced and checked.

17 September 2026: 4 links

2026-09-17T00:00:00Z

Four pieces show that when AI agents do the building, people's judgment moves to either end of the work. People state the problem and write checks before the build, and review the running result before it reaches production.

The Pragmatic Engineer: Inside OpenAI's agentic software factory. Gergely Orosz describes how OpenAI builds software with its coding agent Codex: people define the problem, agents write and review the code, and high-risk changes can require a person's review.
Product Talk: 4 New Evals and 16 Experiment Variants to Fix 1 Customer Complaint. Teresa Torres spent three weeks and sixteen experiments fixing one customer complaint about an AI product, starting by measuring how often the error occurred.
Product Talk: Delivery Isn't Free, All Things Product Podcast. Teresa Torres and Petra Wille argue that AI coding agents made building a feature cheaper but did not make delivering a reliable product free.
Vercel: How Delphi ships 100 times a day with its Python backend on Vercel. Delphi, a company with ten engineers, ships to production over 100 times a day and first sees its AI agents' work on live preview links, according to Vercel.

16 September 2026: 5 links

2026-09-16T00:00:00Z

Five pieces show that when AI helps produce work, a person still has to check it against a source before others rely on it. Where nobody inside the company does that checking, the users end up doing it themselves.

Nielsen Norman Group: AI Can Help Write an Article, but It Can't Stand Behind It. Raluca Budiu, an editor at Nielsen Norman Group, explains that the firm uses AI to write and critique articles, but only a human expert decides whether it publishes them.
Behind the Craft: Stop Building AI Agents. Start Building AI Employees. Pedro Franceschi, chief executive of Brex, says each AI agent should be set up like a new employee, with one job, written instructions, a human manager and a budget.
Nielsen Norman Group: Test Complex Interactions Earlier with AI Prototyping. Megan Chan of Nielsen Norman Group shows designers using AI tools to test complex interfaces with users before engineers build them, and warns that a polished prototype is not finished.
The Product Picnic: AI hypervigilance is now an omnipresent cognitive load for your users. Pavel Samsonov argues that AI has made scams and machine-written content so common that users now check everything they read, and that companies have shifted this effort onto them.
The Beautiful Mess: TBM 439: Day At The Gig. John Cutler, starting a new job, describes how hard it is to learn a product from its notes and maps, and argues that keeping such context current deserves more attention.