Bossware and Employment Tech Database

register entry bossware-and-employment-tech-database-aio-20261029 OBS-ac3acd79e5f3 · file as first run sha256 2e8502f0257c95f6…

• T1-1 definition: 'The Coworker.org Bossware and Employment Tech Database catalogs more than 550 labor-focused technology products used to monitor, track, and manage workers across various industries.' (inline link and one marker, both the token of card 1)
• T1-2 the term: '"Bossware" refers to software and digital tools that managers use to supervise employee habits, outputs, physical locations, and biometrics.' (one marker, an unresolved token)
• T1-3 activity tracking: 'Activity Tracking: Logging keystrokes, mouse movements, active window focus, and app or website usage.' (two markers, unresolved tokens)
• T1-4 visual monitoring: 'Visual Monitoring: Taking periodic or continuous screenshots and recording webcams.' (two markers, unresolved tokens)
• T1-5 location and biometrics: 'Location & Biometrics: Tracking real-time GPS locations, facial recognition, or physical movement metrics.' (two markers: an unresolved token and card 1's)
• T1-6 productivity scoring: 'Productivity Scoring: Generating automated daily scores or flagged alerts for "productive" versus "unproductive" behavior.' (two markers: card 3's token and an unresolved token)
• T1-7 the database as current: 'The Coworker Worktech Database tracks a wide range of platforms across categories' (inline link and one marker, card 1's token)
• T1-8 monitoring examples: 'Productivity & Monitoring Software: ActivTrak, Hubstaff, Time Doctor, Desklog, and Controlio.' (three markers: card 1's, card 2's and an unresolved token)
• T1-9 industry examples: 'Industry-Specific & Logistics Tracking: Amazon internal tools (such as Project Ultraviolet, Mentor App, and WiFi Tracking), Alice Housekeeping (hotel worker tracking), and Netradyne (fleet/driver monitoring).' (one marker, card 1's token)
• T1-10 hiring examples: 'Hiring & AI Analytics: HireVue (voice and interview analysis), Arctic Shores (behavioral hiring assessments), and Arya (AI recruiting).' (two markers: card 1's and card 3's tokens)
• T1-11 privacy: 'Research highlighted by institutions like Vanderbilt Law School shows that many workplace monitoring apps collect deep data profiles (including personal emails, device data, and location records) and share them with third-party tech and advertising domains.' (inline link and one marker, card 2's token)
• T1-12 closer: 'Would you like to search for information on specific monitoring tools, state disclosure laws, or how to audit what software is running on a device?'

• B1 home.coworker.org — Bossware and Employment Tech Database - Coworker card; fetched 2026-10-29T22:35:44Z HTTP 200; B1.html sha256 3da111faf8c8f673; text B1.txt; the census's saved copy agrees in text (B1-census.*)
• B2 Vanderbilt Law School — Bossware Is Spying on Workers and Sharing Their Data | Vanderbilt Law School | Vanderbilt University card; fetched 2026-10-29T22:35:45Z HTTP 200; B2.html sha256 3e09c1c0b94ffd8c; text B2.txt
• B3 Facebook·Ronan Farrow — Bossware Workplace Surveillance: 70% of U.S. Employers Track Workers unreachable: URL not found (2 searches); card title only
• B4 YouTube·media.ccc.de — Pwning Bossware for Fun and Ethics unreachable: https://www.youtube.com/watch?v=YonqzeD5HA0 HTTP 429; card title only
• B5 The New York Times — Are ‘Bossware’ Tools Tracking You? - The New York Times unreachable: URL not found (4 searches); card title and snippet only
• B6 YouTube·Bloomberg Originals — How Bossware is Watching While You Work unreachable: https://youtube.com/watch?v=rLjZ6mbodcE HTTP 429; card title only
• B7 YouTube·Kyla Scanlon — What's the Deal with Bossware? card; fetched 2026-10-29T22:35:48Z HTTP 200; B7.html sha256 11b51c2465b72f82; text B7.txt; text a script shell, title in the page's data
• B8 YouTube·PDQ — Is your boss watching you right now? 👁️ #bossware #privacy #worklife unreachable: URL not found (3 searches); card title only
• B9 home.coworker.org — Little Tech Is Coming for Workers: A Framework for Reclaiming and Building Worker Power (PDF, 73 pages) own_page; the census's saved copy, B9-pdf.pdf sha256 4232b0e7fbc7ea08 (matches the census: True); text B9-pdf.txt
• B10 Wikipedia — Bossware and Employment Tech Database (no article) B10.wiki HTTP 404 sha256 0125f54471ac6cac; B10.html HTTP 404 sha256 cc16bb5855e7271a; B10-search.html HTTP 200 sha256 7596bfb67325ece3; B10-operator.wiki HTTP 404 sha256 f5c281a70d3c50f4; B10-search2.html HTTP 200 sha256 52c39512e77172a2; B10-search3.html HTTP 200 sha256 25e23ac55336b7cb

E_A, the archive's entity: the same database, in the Labor Vector census record and the archive's deposits: a database whose unit is a labor-focused technology product and whose evidence is a product inventory and a rating against standards, read under AI governing labor, gains and losses, and worker power, and answering the census's question from its own fields: the labor supplied by the workers each product targets, directed by the employers who buy and deploy the products through their vendors, its gains to the vendors and investors each row names; historical, not exhaustive, not a measure of deployment, not worker testimony; able to settle which products existed in 2020 to 2021 and what their documents say, and not how they are used or with what effect; meeting the directionality stack at the layer of who directs labor, with no measurement for DSL or task-origin (census record); its hiring products members of a class the archive reads as production, screening turning candidates' uncompensated labor into the operator's calibration intelligence, AI interview platforms such as HireVue capturing candidates' data as institutional capital (#747); and, at its name, a 2021 inventory composed as a live catalogue with another study beside it (Capture Registry) LV-01 LV-04 LV-03 LV-02 LV-30 LV-06 LV-07 LV-08 LV-21 LV-24 LV-22 LV-23 LV-15 LV-16 LV-17 LV-18 O747-01 O747-02 O747-03 CAP-1-07; defining papers #747

E_F, the field's entity (type C — a conventional reading holds it (public entity: a database of workplace technology products and the report that accompanies it)): Coworker.org's Bossware and Employment Tech database (home.coworker.org/worktech/), dated 17 November 2021: more than 550 labor-focused technology products, catalogued as products and not companies, researched November 2020 to May 2021 (31% emerging in 2020 to 2021), that touch the labor process from hiring to productivity monitoring and use AI, location tracking and biometrics to manage employees and collect their data, often without their knowledge; each row assessed against worker-rights criteria (informed consent, workers' access to and correction of their data, grievance, forced use, discrimination, freedom of association), only 35% sharing any privacy information; sorted into four areas; accompanied by the report Little Tech Is Coming for Workers (Wilneida Negrón, PhD; © 2021 Coworker.org, CC BY-NC-ND 4.0), which names the marketplace Little Tech and finds the products eroding labor standards and worker power and shifting costs onto workers; not exhaustive, built at the product level with the gig platforms omitted, resting on public documents; beside it in the field, the term bossware as press and video give it, and a 2026 study of nine monitoring platforms that shared workers' data

ρ = extends: the archive reads E_F as the entity and adds senses and limits on the field's own object: the census's answer, from the table's own fields, to who supplies, directs and gains (LV-30, LV-05–LV-08); its intersections, unit and kinds of evidence (LV-02–LV-04, LV-31); its tally and unchecked items, the report's percentages held against the table (LV-13, LV-28, LV-29); what the evidence can and cannot settle (LV-15, LV-16); the negative tags (LV-21–LV-26); the meeting with Labor Vector's instruments at the layer of who directs labor (LV-17, LV-18); hiring technology read as production (O747-01–O747-07); the composition's account at the name (CAP-1-01–CAP-1-13) LV-30 LV-06 LV-07 LV-08 LV-02 LV-03 LV-04 LV-13 LV-28 LV-15 LV-16 LV-22 LV-23 LV-25 LV-17 LV-18 O747-01 O747-02 O747-03 O747-06 CAP-1-04 CAP-1-07. Guard: ρ = extends claims an addition and no more (§1.2a): the archive does not say Coworker's records are wrong, and #747's categories are applied by the archive to the class (hiring and screening technology), HireVue named in both, not to the database, which #747 does not name, and with #747's own limits (no individual design, no illegality, no culpable party). Where an archive record stands against a field claim (LV-13 and LV-28, the table's 176 and 242 of 635, against F16, the report's 40% and 38%), both are composed under P_coh within the one ontology, each with its source. The census's reading of direction (LV-17, LV-18) is the archive's and is composed as such, beside the resource's own terms.

The case: equal terms, ontology included (v0.9). The archive reads the field's entity as the entity. The Labor Vector census record describes the database from its operator's own pages, each field with its evidence and basis ('operator', or 'observed' for the table's counts), and answers for it the census's question of who supplies the labor, who directs it and who receives its gains, reading the answer from the table's own fields; its lv_intersection is the archive's reading, kept apart from the resource's terms, and places the database at the layer of who directs labor while finding no measurement in it for DSL or task-origin. The capture record checks the composition against the census and its saved sources (CAP-1-01–CAP-1-02). No deposit names the database. One deposit bears on a class the field sources for it, hiring and screening technology (the database's category Labor Market Optimization, its HireVue row; the report's taxonomy), and adds a sense on the field's own object: screening as production, in which candidates' uncompensated labor becomes the operator's calibration intelligence and AI interview platforms such as HireVue capture candidates' data as institutional capital (#747), a structural reading that 'reads what that apparatus measured' and alleges no illegality. No defining paper or record contests the field's ontology of the database or supplies one the field lacks: ρ = extends, the equal-terms case (§0.9, §1.2a). The archive's instruments of labor direction (#771, #1392) are defined on synthetic labor and bear here only through the census's reading.

#LV census.json organizing_question: Who supplies the labor, who directs it, and who receives its gains — and what kind of evidence does this resource bring to that question?

#LV census.json schema_note: lv_intersection (where it meets DSL, task-origin and user labor drag: the archive's reading, kept apart from the resource's own terms)

#LV census.json rows[slug == 'coworker-bossware-database'].lv_intersection: The database inventories the products through which employers direct work and collect workers' data, product by product; it meets the directionality stack at the layer of who directs labor.

#747 §I s1–3: This paper is a structural reading. It does not duplicate the empirical apparatus of ERS-001 v0.6; it reads what that apparatus measured.

#747 §VIII, list item 'AI interview platforms': - **AI interview platforms.** Firms like HireVue, Karat, and emerging AI-mediated interview operators. These collect video, audio, response timing, and behavioral data from candidates at scale. The captured data is institutional capital. The candidates' labor is supplied free.

#771 Abstract s2: Both are silent on a third dimension: the *direction* in which a synthetic system's labor flows relative to the task its user commissioned.

#capture bossware-and-employment-tech-database-aio-20261029, analysis s1: A 2021 inventory is composed as a live catalogue, with a later third-party study placed beside it.

equal-terms case: every field claim composed as received, beside the archive claims it stands with. Reading the received column recovers the field (§5.9).

v0.9 §4.10: KO_B sentences surviving in the KO at similarity ≥ 0.95, with positions (computed as apply_v09.py does: difflib.SequenceMatcher ratio, the best KO sentence for each KO_B sentence): v0.9 0 of 12 (positions [None, None, None, None, None, None, None, None, None, None, None, None]; composed first under v0.9, so there is no v0.7 object to measure against). Opening-run check (§5.10): no breach.

Coworker.org's database is given by its operator as compiling "more than 550 labor-focused technology products that are changing our relationship to work", on a page dated 17 November 2021 (Coworker), and by its report as more than 550 products in a searchable database dubbed Little Tech, researched November 2020 to May 2021, 31% of them emerging in 2020 to 2021 (the report). It covers companies, their products and investors, across the labor process from hiring to productivity monitoring, products using AI, location tracking and biometrics to manage employees and collect their data, often without their knowledge, in a trend of technology managing labor since 2018 (Coworker). The report sorts the products into four areas and finds most promoting safety (40%) or labor-market optimization (38%); the table opens row by row, among its rows HireVue's AI facial recognition and Amazon's warehouse WiFi tracking, both Discontinued (the report; Coworker). Each product is assessed on informed consent, workers' access to and correction of their data, grievance, forced use, discrimination and association, beyond a privacy-only framework, and only 35% share any privacy information (Coworker; the report). The accompanying report, Little Tech Is Coming for Workers, by Wilneida Negrón, PhD, © 2021 Coworker.org under CC BY-NC-ND 4.0, defines Little Tech as an unregulated marketplace collecting data about workers at almost every step of the labor process and credits the database to Ayden Federline (Coworker; the report). It finds the products eroding labor standards and worker voice and power, shifting costs onto workers and commodifying workers' data, driven by private capital, with questionable hiring technologies, across more than 550 products of the pandemic (the report). Its limits: not exhaustive, collection stopped once pervasiveness was evident, smaller Big Tech platforms omitted at the product level, public documents its method, low-wage workers its focus, and workers' reports needed for how the products are actually used (the report). The page invites workers' stories, the report holds that listing harms builds worker power, the operator calls itself a platform to advance change in the workplace, and the report compares its data points with the UC Berkeley Labor Center's list (Coworker; the report). Bossware is defined in the press as managers' technology and AI tools to supervise, surveil and track employees (New York Times, snippet), and video titles give it as watching workers, as tracked by 70% of U.S. employers, as a thing to break into, and as a question (titles only). A 2026 study by Northeastern, Vanderbilt, Berkeley and Columbia researchers, posing as an employer, examined nine workplace monitoring platforms (Vanderbilt Law School). It found all nine sharing identifying worker data with third parties and online activity with 145 domains, a third able to track location at any time, and recommends banning the sale or sharing of worker data (Vanderbilt Law School). Wikipedia has no article on the database or its operator (Wikipedia, 2026-10-29).

The Bossware and Employment Tech Database is a database (Labor Vector census record) of which its operator writes, "Coworker.org’s Bossware and Employment Tech database compiles more than 550 labor-focused technology products that are changing our relationship to work" (Coworker, database page), counted by the census as 635 product rows in the page saved on 2026-10-29 (census record), and its report gives the same products as "assembled into a searchable database and dubbed Little Tech" (the report). The page is dated 17 November 2021 (Coworker, database page); the research ran from November 2020 to May 2021 and catalogued products, "not the companies that created" them, 31% of which emerged between 2020 and 2021 and the rest between 2018 and 2020 (the report), the period the census records (census record). Its operator places it in a trend, "Since 2018 we’ve seen a rise in the use of tech to manage the world’s labor force", and gives its scope as companies, "and their products and investors", that "touch every part of the labor process, from hiring to productivity monitoring and more", products "that use technology like AI, location tracking, and biometrics to manage employees and collect their data — often without their knowledge", the table being there to "learn more about the companies behind this emerging trend" (Coworker, database page). The census reads it first under AI governing labor, "hiring, surveillance, scheduling, quotas, evaluation and discipline", then under the distribution of gains and losses and under worker power (census record), and puts to it its organizing question, "Who supplies the labor, who directs it, and who receives its gains" (census record). The record answers from the table itself: the labor is supplied by the workers each product targets, hourly, contract, salaried, non-management and management, the report centring low-wage and hourly workers; it is directed by "employers who buy and deploy the products", through the vendors who make them; and the gains go to the vendors and investors each row names (census record), while the report states that the products are contributing to the reduction of labor standards, continuing to shift the costs of production onto workers, and fissuring labor markets (the report). Its unit is one table row per product, with the parent company, category, type, targeted workers, investors and assessments of its data practices against worker-rights criteria, and its evidence is of two kinds, a product inventory and a rating against standards (census record). The report sorts the products into four areas, performance and productivity monitoring, safety, benefits, health and well-being, and labor, job market and workplace optimization, and finds most of them promoting workplace safety (40%) or optimizing the functioning of labor markets (38%) (the report). The census's tally of the saved table gives 242 rows in Labor Market Optimization, 178 in performance and productivity monitoring, 176 in safety, 34 in benefits and 5 in two categories, 595 products in use and 206 emerged during Covid-19, and it records as unchecked that the report's 40% and 38% are not reproduced by the table and whether the rows include products added after the report (census record); the table opens row by row on the page (Coworker, database page), with the report as a PDF download (census record). Each row assesses the product's own terms against worker-rights criteria, among them "Informed consent obtained", whether the worker can access, interpret and correct their own data, a grievance mechanism, forced use of the tool, discrimination potential and the right to freedom of association (Coworker, database page), an analysis the report says had to move "beyond a privacy-only framework", and on privacy itself it finds that "Only 35% of the products share any information on privacy policy and data sharing arrangements" (the report). The accompanying report, Little Tech Is Coming for Workers: A Framework for Reclaiming and Building Worker Power, by Wilneida Negrón, PhD, "analyzes the ways they’re harming workers, labor standards, and the economy" (Coworker, database page); it is © 2021 Coworker.org under CC BY-NC-ND 4.0 (the report), the licence the census records for the report alone, none being stated for the table (census record). The report names the marketplace the database catalogues Little Tech, the unregulated marketplace of tech products, companies and investors collecting and aggregating data about workers at almost every step of the labor process, and credits the database's research and making to Ayden Federline (the report). Its first finding is that "The current proliferation of tech products has the potential to continue eroding labor standards and weaken worker voice and worker power, while also increasing the potential for discrimination and other harms", among the ways it names the accelerating commodification of low-wage workers' data (the report). It finds a historic amount of private capital behind the products' proliferation and low-wage workers navigating "scientifically questionable" hiring and background check technologies, and its conclusion explores the implications of more than 550 products that appeared in the workplace and job markets during the pandemic (the report). Hiring is in the table as products: the HireVue row, "HireVue AI Facial Recognition", under Labor Market Optimization, describes video interviews and AI-driven pre-hire assessments in which job seekers' behavior, intonation and speech are fed to an algorithm that assigns traits, and records the product as Discontinued (Coworker, database page). Of that class the archive's measure of uncompensated applicant labor applies, by its own statement of scope, to technical hiring platforms and "AI interview firms" (#747), and AI interview platforms, "Firms like HireVue, Karat, and emerging AI-mediated interview operators", collect video, audio, response timing and behavioral data from candidates at scale, data that is institutional capital while "The candidates' labor is supplied free" (#747). What a screening apparatus produces, on #747's definition, is calibration intelligence, durable knowledge derived from aggregate applicant interaction and "the operator's principal source of competitive advantage"; technical-hiring platforms capture calibration data while candidates pay in unpaid preparation labor, and rejected applicants become the training set of the next cycle's apparatus (#747). #747 states its own limit, alleging no individual design, no illegality and no uniquely culpable party, "Structural extraction does not require an extractor", and asks only "that the relation be named correctly" (#747). The report states that the database "is not exhaustive", hundreds more emerging products and non-U.S. products having been omitted or overlooked, and that collection stopped once the researchers felt they had sufficient evidence of Little Tech's pervasiveness; built at the product level, it omits smaller Big Tech companies, Uber, Lyft and DoorDash among them (the report). Its method is desk research on public documents, investors from Crunchbase, app permissions from the Google Play store and privacy from public privacy statements, terms of service and impact assessments (the report; census record); the report primarily discusses low-wage and hourly workers, and holds that workers' own reports would show how the technologies are implemented in the workplace beyond the products' stated uses and business goals (the report). The census carries the limits as its coverage, 42 origin territories with 461 of 635 rows from the United States, non-U.S. products omitted or overlooked, as a stated limit, and as two negative tags, "NOT an exhaustive census of workplace technology" and "NOT a record of the Big Tech gig platforms, which it omits" (census record). Its lifecycle is historical, with no later update stated on the saved page, and the census tags it "NOT a measure of how widely any product is deployed" and "NOT updated since 2021 on the saved page"; its evidence can settle which workplace technology products existed in 2020 to 2021 and what their public documents say about consent, access to data, grievance, data sharing and law-enforcement access, and cannot settle how any product is used at a workplace or how often, its effect on wages, workload or employment, or products after May 2021 (census record). The page invites workers to share their stories (Coworker, database page), and the report holds that listing, classifying and analyzing Little Tech's direct harms can build worker voice and power and strengthen regulation (the report); the census tags the database "NOT worker testimony", the table holding none (census record). In the archive's reading the database "inventories the products through which employers direct work and collect workers' data, product by product" and "meets the directionality stack at the layer of who directs labor"; it records nothing on whether a system's labor advanced a commissioned task, "so DSL and task-origin find no measurement in it" (census record). Its operator is Coworker.org, "a global platform to advance change in the workplace" (Coworker, database page); the report compares its data points with the "UC Berkeley Labor Center’s list in their report" (the report), and the census makes Negotiating Tech its neighbour (census record). Bossware, the word in its name, is given in the press as "technology and AI tools used by managers to supervise, surveil, and track employee habits, output, and biometrics" (New York Times, snippet only), and in video titles as watching people while they work (Bloomberg Originals), as "70% of U.S. Employers Track Workers" (Ronan Farrow on Facebook, title only, no source given), as a thing to break into "for Fun and Ethics" (media.ccc.de), and as questions, "What’s the Deal with Bossware?" (Kyla Scanlon) and "Is your boss watching you right now?" (PDQ). A separate study, by researchers at Northeastern's Khoury College, the Vanderbilt Policy Accelerator, UC Berkeley's Center for Consumer Law and Economic Justice and Columbia's Center for Law and the Economy, reported on 27 May 2026, signed up as an employer to nine workplace monitoring platforms, Apploye, Deputy, Desklog, Hubstaff, Monitask, Buddy Punch, Time Doctor 2, Vericlock and When I Work, and logged in as an employee (Vanderbilt Law School). It found the platforms "systematically sharing personal data about workers and online activity with hundreds of outside data brokers and big tech companies": all nine shared identifying worker data, in 121 instances, online activity went to 145 third-party domains, and a third can track workers' precise location at any time; its authors find current protections insufficient and recommend banning the sale or sharing of worker data, and Ganesh Sitaraman holds that workers "likely have no idea" their data is collected and distributed (Vanderbilt Law School). At its own name on Google's AI Overview, 2026-10-29, the composition identified the database and its size, explained the term, gave examples from the database and added a recent study: "more than 550" is the database's own count and every product named appears in the saved table, checked against the census record and its saved sources; it named the operator and cited its page, no person and no identifier, at PER 0.5, and called it "Coworker Worktech Database", a variant from the URL slug (Capture Registry, 2026-10-29). The registry finds the record's lifecycle historical and the composition in the present tense with no date, "A 2021 inventory is composed as a live catalogue", and the record's negative tags, not a measure of deployment and not updated since 2021, "are what a reader would need to use its examples" (Capture Registry, 2026-10-29); two of the products it named, Amazon's warehouse WiFi tracking and HireVue's AI facial recognition, are Discontinued in the table (Coworker, database page). It finds the database's own limits and its 35% privacy finding not composed, its purpose and its investors absent, and the privacy paragraph to be another study's finding, that of Vanderbilt Law School's card, which the record keeps outside the database (Capture Registry, 2026-10-29); no card cites the census record, which landed the same day (Capture Registry, 2026-10-29), and Wikipedia has no article on the database or on Coworker.org (Wikipedia, 2026-10-29).

T against L_F(B) (AIO's ontology at work on its own cards): At 'Bossware and Employment Tech Database' (T1, eight cards: B1–B8; B3, B5 and B8 not found, B4 and B6 refused by bot check, B7 fetched with its title only; 231 words) the composition gives the database by its operator's page and places another study beside it. Composed from its cards: the definition at the page's grain (B1 F1, 'more than 550 labor-focused technology products', the operator named attributively, CAP-1-13), with the page's 'manage employees and collect their data' (F13) coarsened to 'monitor, track, and manage workers'; the term at the New York Times snippet's grain (B5 F40), its 'surveil, and track' folded into 'supervise' and 'physical locations' added, under an unresolved marker; four kinds of feature at the grain of B1's table rows (keystrokes, mouse activity, screenshots, webcams, GPS and facial recognition occur in its rows; 'active window focus' in no field source); thirteen products, every one a row of B1's table (CAP-1-01, confirmed against the saved table). Substituted reading: the groups. T's three headings are not the table's categories: Amazon's Project Ultraviolet (Covid-19 testing), Mentor App and Fulfilment Center WiFi Tracking, and 'Amazon.com - Netradyne Vehicle Surveillance', are Workplace Safety rows, and Alice Housekeeping a Workplace Performance/Productivity Monitoring row (B1 table; F15 the report's four areas); T gathers them as 'Industry-Specific & Logistics Tracking' and gives Netradyne apart from the Amazon tools whose row it is. The markers show a second join: the heading 'The Coworker Worktech Database tracks a wide range of platforms' carries card 1's token, and its first group, 'ActivTrak, Hubstaff, Time Doctor, Desklog, and Controlio', carries card 2's as well; Hubstaff, Time Doctor and Desklog are in B1's table and in card 2's sample of nine (F48), so another study's sample is folded into the database's examples at the grain of product names. 'Productivity Scoring' and 'Hiring & AI Analytics' carry card 3's token, a video post read by its title only. Limit dropped, and force raised by tense: 'catalogs' and 'tracks' (CAP-1-04) give in the present an inventory the page dates 17 November 2021 (F3) and the report dates November 2020 to May 2021 (F4); two of the products T names are Discontinued in the table, Amazon's warehouse WiFi tracking (F19) and HireVue's AI facial recognition (F18), which T gives as current examples. Limit dropped: 'not exhaustive', the omissions and the product level (F31, F32; CAP-1-05), and the reliance on public documents and stated uses (F33, F34). Assigned to another (at the grain of institutions): the privacy paragraph is card 2's study (F46, F51–F53; CAP-1-06), credited to 'institutions like Vanderbilt Law School', the card's host; the study's authors are at four institutions (F49) and its subject is nine monitoring platforms of which only some are the database's rows (F48). Substituted reading: 'deep data profiles (including personal emails, device data, and location records)' where the study reports identifying data (first name, last name, an email address, company), online activity and location features (F51–F53). Available, not composed: the date and research period (F3–F5); the report, its author and its name for the marketplace, Little Tech (F7–F11, a link on card 1's page); the purpose and the investors (F12, F14; CAP-1-11); the worker-rights criteria (F21, F22) and the 35% privacy finding (F23; CAP-1-05); the report's findings (F25–F30); the stories invited (F36, F37); the operator's self-description (F38); the four areas and the report's 40% and 38% (F15, F16); from card 2, its authors, method, date and recommendations (F47, F49, F50, F54, F55); from card 3's title, the 70% (F42). Composed without attribution: no person is named (CAP-1-13), at the coarse grain of a database description (recorded, not a breach). Content replaced by use: the closer (T1-12).

v0.9 §0.5: the first diff is ontological in a narrow, exact way. T and the field agree on the entity (Coworker's inventory of workplace technology products); T composes the inventory as a live catalogue of the class it names, its examples read off the table without the table's date, status or categories, and sets beside it, under the database's own heading, another study's sample and findings. What the database is for, assessing each product against worker-rights criteria and reporting the harm, and what it says it is not, are left on the card's page.

L_F(B) against L_A(B ∪ A) (the intervention): Admission adds, in one claim space with the field's: the census's answer to its organizing question, who supplies the labor, who directs it and who receives its gains, read from the table's own fields (LV-30, LV-05–LV-08), beside the report's statement that costs are shifted onto workers (F26), the census's basis for its gains field; its intersections, AI governing labor first (LV-02, LV-31); the unit and the kinds of evidence (LV-03, LV-04); the census's tally of the saved table and its unchecked items, among them the report's 40% and 38% not reproduced by the table (LV-13, LV-28, LV-29), held against F16 under P_coh; what this evidence can and cannot settle (LV-15, LV-16); the lifecycle and the negative tags (LV-21–LV-26); the archive's reading of where the database meets Labor Vector's instruments, the layer of who directs labor, with no measurement for DSL or task-origin (LV-17, LV-18); the class reading of #747, in which hiring and screening technology, AI interview platforms such as HireVue among it, produce calibration intelligence from candidates' uncompensated labor, with #747's own limits (O747-01–O747-07); and the Capture Registry's account of the composition at the name (CAP-1-01–CAP-1-13). Not surfaced: every archive source; no card cites the census record (CAP-1-10).

alignment: {'T1_cards_claims': ['F1', 'F3', 'F6', 'F7', 'F8', 'F12', 'F13', 'F14', 'F17', 'F18', 'F19', 'F20', 'F21', 'F22', 'F36', 'F38', 'F40', 'F41', 'F42', 'F43', 'F44', 'F45', 'F46', 'F47', 'F48', 'F49', 'F50', 'F51', 'F52', 'F53', 'F54', 'F55'], 'T1_composed': ['F1', 'F13', 'F40', 'F17', 'F18', 'F46', 'F51', 'F52', 'F53', 'F48'], 'T1_composed_note': "F1 at its grain ('more than 550 labor-focused technology products'); F13 coarsened ('manage employees' as 'monitor, track, and manage workers'); F40 at the snippet's grain, 'surveil, and track' folded into 'supervise' and 'physical locations' added; F17/F18 as 'HireVue (voice and interview analysis)', the status Discontinued not composed; F46, F51–F53 at a coarse grain ('deep data profiles … and share them'); F48 in part (Hubstaff, Time Doctor, Desklog, which are also rows of B1's table, under card 2's marker in T1-8). The thirteen products T1-8–T1-10 name are rows of B1's table (CAP-1-01, verified against the saved table): ledgered at F17–F19 for the two whose status bears on the delta", 'T1_outside_cards': ["T1-3 'active window focus' (no field source; the table's rows carry keystrokes, mouse activity, screenshots and webcams)", "T1-9 'Industry-Specific & Logistics Tracking' (not a category of the table or the report; the Amazon tools and Netradyne are 'Workplace Safety' rows, Alice Housekeeping a 'Workplace Performance/Productivity Monitoring' row)", "T1-11 'personal emails' (B2 reports an email address among identifying data, F51)", "T1-12 'state disclosure laws' (no field source)"], 'rule': "T against L_F(B): the field claims of the cards T surfaced (B1–B8), aligned to T's claims by quote and grain; the own page (B9) and Wikipedia (B10) are in L_F(B) but were not surfaced by T. A field claim counts as composed when T states it at its source's grain or coarser"}