--- title: "Practitioner's Corner" canonical: "https://current.tinyfish.ai/issue/52/practitioners-corner" publication: "TinyFish" publisher: "TinyFish" language: "en-US" guide: "https://current.tinyfish.ai/llms.txt" permalink: "https://current.tinyfish.ai/issue/52/practitioners-corner" issue: 52 section: "practitioners-corner" description: "Lessons from the field—what we see building at scale" authors: ["Sevda Polat","Sable Whitford"] published: 2026-10-26 updated: 2026-10-24 articles: 14 --- # TinyFish > CURRENT is a journal about AI agents in production, for the engineering and product leaders who have to make them work. Drawing on experience building enterprise web agents, it explains why the demo succeeded and the deployment didn't, and what that means for choosing vendors, briefing the board, and avoiding mistakes others have already made. [Read on the website]() ## Issue 52 — Practitioner's Corner Lessons from the field—what we see building at scale ### Builder Profile - [Ananya Gubbi Mohanbabu and the Control You Can Get Back To]() — By [Sevda Polat]() · Published 2026-10-23 · Format: In Practice Ananya Gubbi Mohanbabu, a UC Berkeley PhD student, builds tools for blind and low-vision people who use screen readers, software that speaks the screen aloud. One of her systems hides page elements that seem irrelevant to the user's stated goal. In a 12-person study, its screen reader users finished tasks in about half the time they took without it. It can also hide the button someone needed. Her newest work tests AI agents that click and type for you, and asks what a person who cannot watch the screen needs in order to follow one, stop it, and take over. ### Feature Essay - [The Questions Completion Rates Never See]() — By [Sevda Polat]() · Published 2026-10-23 · Format: Feature Essay Boston gave residents seven categories for reporting city problems, plus "other." When a researcher read a sample of reports, nearly half sat in "other," and trash, among the most common complaints, had no box of its own. The city could see that gap only because people still reached for "other." ### Imagined Interview - [Your Ear Is Not a Lock, and Other Things a Pharmacy Technician Told Us Over the Phone]() — Published 2026-10-23 Every pharmacy phone line is a small courtroom that convenes forty times a day without a bailiff. A voice says it's calling about somebody's metformin. Behind the counter sits a technician, usually not the pharmacist, with a label printer chirping at her elbow. She has a few seconds to rule on two questions. Who is this? And what are they allowed to hear? ### Short Explainer - [When a Payment Request Gets No Answer]() — By [Sable Whitford]() · Published 2026-10-23 Say a program asks a server to make a payment, and nothing comes back. No error, no receipt. The connection hangs for a while, then gives up. On the server, the payment might have gone through, might be half done, or might never have started, and the program can't tell which. ### Further Reading - **Sorting Things Out: Classification and Its Consequences** — [Source]() — Published 2026-10-23 Bowker and Star study disease codes, nursing records, and apartheid's race categories, showing what each scheme counts and what it drops. - **Questions and Answers in Attitude Surveys: Experiments on Question Form, Wording, and Context** — [Source]() — Published 2026-10-23 Schuman and Presser show that wording, question order, and offering "don't know" all change answers. Open and closed questions mislead differently. - **The Power to Name: Representation in Library Catalogs** — [Source]() — Published 2026-10-23 Olson shows how subject headings carry their era's exclusions, so searchers can only find what the catalog has words for. - **Kinds of People: Moving Targets** — [Source]() — Published 2026-10-23 Hacking's "looping effect": people who are classified start describing themselves in the category's terms, and the category shifts in response. - **PhantomFill: What Language Models Do With a Required Field** — [Source]() — Published 2026-10-23 A 2026 preprint retests the survey "don't know" problem on models. Make a field mandatory, and they invent values instead of reporting none. - **FormFactory: A Benchmark for Agents That Fill Out Web Forms** — [Source]() — Published 2026-10-23 Scores agents separately on picking the right value and on entering it through dropdowns, date pickers, and checkboxes. ### Past Articles - [How Phil Cuvin Made Agent Manipulation Measurable]() — By [Sevda Polat]() · Published 2026-10-16 · Format: In Practice A web agent that reads an interface well enough to book your flight can also read it well enough to be fooled by a fake "required" checkbox or a "most popular" badge. Phil Cuvin, along with co-authors Hao Zhu and Diyi Yang, built a benchmark called DECEPTION to measure how often this happens under controlled conditions. Agents fell for common manipulation tricks in over 70% of test tasks; humans averaged 31%. The vulnerability profiles were nearly inverted: agents are most susceptible at the points where they read the interface most carefully. - [Variance Is the Constraint]() — By [Sevda Polat]() · Published 2026-10-16 · Format: Feature Essay For decades, U.S. states reported average highway travel times. Two routes across New Jersey might average 58 and 59 minutes, functionally the same. But one swung nearly twice as much from day to day. The averages said the routes were interchangeable; the variance said they weren't. Federal rules eventually required states to report travel-time *reliability* alongside the mean. Web agents have the same measurement problem. A recent study ran the same agents on the same tasks three times: 46.5% succeeded at least once, but only 27.8% succeeded every time. - [What Can the Next Person Still Do?]() — By [Sable Whitford]() · Published 2026-10-09 · Format: Central Essay Norwegian conservators saved a thousand-year-old Viking ship by soaking it in hot alum. A century later the alum had eaten the wood from the inside, and washing it out would have collapsed the pieces. Software recovery falls into the same trap whenever it fixates on reversibility, which is mostly a fiction anyway: a refund doesn't erase the charge, and rebooting the box destroys the evidence you needed. Conservation ethics asks a better question. After you act, what can the next person still do? - [Jiacheng Miao — Making Papers Callable, Then Testing What That Preserves]() — By [Sevda Polat]() · Published 2026-10-09 · Format: In Practice Jiacheng Miao built a system that turns computational research papers into functions an AI agent can call; 74 of 100 papers converted with no human intervention. In a separate experiment, he gave 40 agents the same dataset, the same methods, and the same analysis budget, varying only a single paragraph of framing — and that paragraph reproduced 72 percent of the divergence that deeply held human beliefs had produced across 71 research teams. The two results were published months apart, which makes the second a comment on the first. ## Other sections in Issue 52 - [Daily Brief]() - [Market Pulse]() - [Foundations]() - [Echoes]() - [Vision]() ## More - [Publication guide]() - [Issue archive]()