Search the Practicum

Type two letters · ↑↓ select · Enter open · Esc close

Type at least two letters.

What is new in the Practicum

A reverse chronological record of material changes. Living pages also show their own verification dates, which matter more when facts change quickly. Current Practicum version: v0.17.1 (2026-08).

2026-08-18 — a full check of the live layer, and a chapter index

  • Every live page was checked against its primary sources. The pass caught four discrepancies that were already misleading readers: DeepSeek V4 Pro pricing was more than twice out of date, and the vendor had introduced time-of-day pricing; OpenAI's in-chat checkout had been withdrawn in March, while the page still presented it as a live product; the EU AI Act's transparency duties came into force on Aug 2, five days after the legal page was last checked; and AI adoption figures were attributed to the wrong McKinsey report.
  • New: the Practicum, chapter by chapter. A reader arriving from a printed chapter used to land on the homepage and hunt. Every chapter of all three books now has a direct route to the material that continues it.
  • The tool map now covers more of the world: Microsoft Copilot, Mistral (renamed Vibe in May), Grok, and a new section on coding agents. Russian-market tools were removed from the English edition and replaced with global equivalents rather than left as gaps.
  • What the live layer turned up: more than half of crawler requests are now for AI training, and a paid-access market has formed around that; the AI infrastructure bottleneck moved from chips to transformers and grid connections; on the ARC-AGI-3 benchmark people solve 100% of environments while frontier models score under 1%.
  • All 13 role pages were deepened with current data, from AI in accounting being adopted as an answer to a staffing shortage, to new mandatory requirements for AI in hiring.

2026-08, v0.17: the full English Practicum clone is ready for author review

  • Every one of the 186 Russian routes now has a functional English path: 184 standalone English pages and two documented legacy redirects to English digest issues. All 11 browser tools have an English interface while keeping localStorage, export, print, and offline behavior.
  • English follows the Russian source as a close functional clone. Each item passed semantic comparison, a humanizer pass, and a post-humanizer recheck. General prose stays at B2; C1 appears only for precise specialist terms, with a plain-language explanation.
  • English search, links, 404 pages, HTML metadata, canonical pairs, and the local language switch have been checked. Automatic country and IP selection remains off until every localization-ledger record has individual author acceptance.
  • The English "Book in an Hour" and Method Map now have HTML and PDF pairs. Two missing Volume 2 author illustrations were added to the source assets so the English static pages do not return 404.
  • v0.17 is live as a technically complete English surface for direct access and author review. This is not final editorial acceptance: all 186 pages and 11 tools are still ready-for-author-review, not author-accepted.
  • On Aug 11, 2026, the author gave a separate explicit instruction to enable automatic geo routing while the maintenance gates remain open. At /, an explicit ?lang= choice and the saved locale win first, followed by browser language and country. The site does not store the IP or send it to an external service. This rollout does not replace author acceptance, freshness, or independent source verification.

2026-08, v0.16: Volume 3, a local Trust Compass, and a safe EN foundation

  • The Volume 3 Workbook gained "My Trust Protocol on One Page": high-cost situations at home and at work, a second channel, independent evidence, a human owner, repair after an error, a 30-day move, and a review date. The local Trust Compass can assemble and print it.
  • A clearly labeled composite family and work example was added. The compass does not judge whether a voice or file is authentic and does not promise to detect a deepfake.
  • The English surface received workbook cores for Volumes 1-3, the Trust Compass, My Profile, separate search, and a language switch. Automatic country selection remains off until full English parity and author acceptance. A partial translation is not presented as a finished product.
  • English defaults to B2, uses C1 only for necessary technical terms, and requires a semantic review, humanizer pass, and another semantic review in the localization ledger.

2026-08, v0.15.8: from analysis to a working procedure

  • Every published RU/EN digest issue from 001 to 005 now follows the Working Map with one or two specific Practicum worksheets. The path leads from a verified change to an action.
  • Issue 004 continues its agent-boundary analysis with the Agent Contract and evaluation set. Issue 005 starts AI patch work with the Human Review Matrix and the same contract.
  • Publication checks require this block between the decision map and the general next step. Sources, historical windows, checksums, and conclusions did not change.
  • Every row in the Working Map for issues 001-005 now links directly to the online article or primary material behind the decision.
  • In 004, "buy or build" links to the original Curative case so readers can separate the case from the Practicum frame.
  • Validation blocks a release pair when a row has no direct external source. Facts and conclusions did not change.

2026-08, v0.15.6: from news to a working decision

  • The last paragraph of every signal in issues 001-005 now combines the fact, direct source, evidence boundary, and a usable move in natural prose.
  • Each issue ends with a bilingual Working Map with two to four decisions, context, first move, and boundary. It supports a choice without presenting a conditional benefit as guaranteed.
  • Issue 004 limits its two-week vibe-coded prototype test to a narrow, reversible process with synthetic data. It does not approve replacement in a regulated or critical process.

2026-08, v0.15.5: every signal ends with a practical move

  • Each short signal in issues 001-005 now ends with its own conditional test, decision, or review for a person, team, or business.
  • The buy-versus-build case in 004 compares a product with a narrow prototype using data, integrations, support, ownership, and rollback. It does not claim that custom code is always cheaper.
  • Sources, facts, historical windows, and the overall conclusions did not change.

2026-08, v0.15.4: a source beside every digest signal

  • Every verifiable signal in published issues 001-005 now has a direct external link beside the claim.
  • This was a citation revision. Wording, conclusions, historical windows, and source-bundle checksums did not change. Issues 002-005 show a short r2 correction note.
  • Publication checks require the same rule for future issues. Secondary reporting and community signals must be labeled and cannot stand as the only evidence for a fact.

2026-08, v0.15.3: archive issues 004 and 005 published

  • Bilingual archive issues 004 and 005 were published with one main analysis, 12 short signals, limitations, and sources.
  • Both show the historical weekly window and the actual publication date, August 7, 2026. Their RU/EN pairs share the same revision and source-bundle checksum.
  • Denis Ermilov accepted the fact check and language review before promotion. Automation for 006 and the daily Reddit Compass remained off.

2026-08, v0.15.2: a new editorial digest format

  • The digest landing page now explains the new format: one verified shift, 10-15 short signals, and three practical areas, people, business, and trust. Issues 001-003 remain a historical archive.
  • The main analysis asks what changed, what it costs or makes valuable, what it means for a person, which business decision follows, and how trust can be checked.
  • Editorial intake accepts several independent sources. Reddit Compass is a labeled discovery source, not the only proof of a fact. Only accepted RU/EN pairs can be published.

2026-07, v0.15.1: Volume 3 source notes clarified

  • Volume 3 source card 79 now carries the required legal note beside the WhatsApp reference. The C2PA card matches the accepted wording in Chapter 7.

2026-07, v0.15: Volume 3 without practical gaps

2026-07, v0.14: evidence and Volume 3 Chapter 10

2026-07, v0.13.3: Volume 3 Chapter 7 accepted

  • The short Volume 3 edition gained the one-page Chapter 7, "Sign the Authentic or Catch the Fake." Its HTML and PDF proof now covered From the Author and Chapters 1-7.
  • Sixteen Chapter 7 source cards and the glossary entries "Passkey" and "Personal Provenance" were added. Volumes 1 and 2 were not rebuilt.

2026-07, v0.13.2: sources checked again

  • Source pages for Volumes 1-3 were rebuilt from one bank after a network check of 514 direct links.
  • Two dead addresses were removed, and Open to Work gained a stable Google Books record. A DOI is added only after the author, title, and year match exactly.
  • Print, digital, and online sources were synchronized. Access limits on one site are not treated as evidence that the source disappeared.

2026-07, v0.13.1: a clear path to the weekly brief

  • The Practicum home still leads to the books and tools. The Weekly Field Brief entry now offers two clear actions: get the next issue or read an example first.
  • RU and EN pages promise the same verified shift, separate fact from interpretation, and state that a person performs editorial review.
  • Issues 001-003 use consecutive weekly dates, July 7, 14, and 21. Basic view statistics were added without paid custom events or another form on the home page.

2026-07, v0.13: the working Volume 3 section, Trust

2026-07, v0.12.1: Volume 3 Chapter 2 accepted

  • "Counterfeit Presence" moved from draft to author-accepted status in the short Volume 3 edition. Its one-page version and PDF were rebuilt without mixing them with Volumes 1 and 2.
  • Sources, glossary, and print proof were synchronized. Chapters 3 and 4 remained drafts, so accepting Chapter 2 did not make the full volume ready.

2026-07, v0.12: a separate short proof for Volume 3

  • The One-Hour Book gained a separate HTML/PDF edition of the current Volume 3 material.
  • Status was explicit: From the Author and Chapter 1 were accepted, while Chapters 2-4 remained review drafts. Their presence did not mean that the volume was accepted.

2026-07, v0.11.2: how to read source cards

  • Source pages for all three volumes gained one reader guide. It explains what a card supports, how links work, and why note numbers may have gaps.
  • Cards connect by volume, chapter, and note number rather than a page number that changes with layout.
  • "Primary" and "additional" show navigation order, not an automatic academic ranking.
  • Source cards now show full reader descriptions and only exact named links. Technical missing-URL messages were removed.
  • Wrong and outdated addresses were corrected and given verification dates. A card without an exact URL remains a bibliographic record instead of receiving a similar but unrelated link.

2026-07, v0.11: clickable sources for three volumes

  • Every volume gained a direct-link page whose numbers match the printed source cards.
  • Print keeps full bibliographic descriptions without long URLs. A QR code leads to the living volume page. Digital editions use the same data and show named active links.

2026-07, v0.10.1: stable search updates

  • Old results remain visible while the next search query runs. The list changes only when the new result is ready, which stops flicker and dialog height jumps.
  • The UI smoke test now covers fast query refinement.

2026-07, v0.10: local content search

  • Site B gained local search across 151 public Practicum pages and their sections, built into a static index.
  • Search opens from the desktop header, Cmd/Ctrl+K, or mobile menu. Orama runs in the browser after two characters, so queries and index content do not go to an outside service.
  • At that release, the UI supported RU and EN while the indexed content remained Russian. Public utility pages, author pages, changelog, English overview, and the One-Hour Book were excluded.

2026-07, v0.9.6: a contact line instead of a CTA tile

  • Telegram and author email moved from large CTA tiles to a compact labeled contact line. The email appeared only after a button action, followed by a clear "write" step.
  • RU and EN heroes named the two-volume series without repeating each volume's purpose beside buttons.

2026-07, v0.9.5: author contacts and a cleaner hero

  • The full author page gained direct Telegram and click-to-reveal email. The server HTML and ordinary mailto: links did not contain the address. This reduced bulk scraping but was not presented as protection from targeted spam.
  • Print directed readers to request the email through Telegram. The RU and EN heroes removed repeated volume descriptions.

2026-07, v0.9.4: temporary English cover

  • During external validation, Site B showed an English editorial home at /. Russian remained at /?lang=ru, and all Russian content routes stayed in place.
  • English page chrome, author dossier, navigation, and footer appeared only on that cover. The temporary exception was documented for a later return to a Russian-first root.

2026-07, v0.9.3: language polish

  • Canonical Russian author text replaced a stiff phrase for "now Emergn" with a more natural one. The sage manifesto kept its unbroken composition as a reading pause.

2026-07, v0.9.2: the author's international start

  • The author story now describes the start of his career at Exigen Services, now Emergn: its global top-four position at the time, US clients, a distributed English-speaking team, and investment projects in the US and Europe.
  • Four proof markers became numbered mini-cards. Canonical RU/EN wording moved to meta/author-profile.md for reuse.

2026-07, v0.9.1: the author and the canonical Volume 2 title

  • The author path now connects Exigen Services, the Bank of Russia, international payments, large technology programs, and public AI-native systems on GitHub.
  • Core pages use the book title, Business in the Age of AI Agents. Navigation keeps the short label "Volume 2."
  • The home, author pages, method map, tool catalog, and Volume 2 Workbook were updated without route or design changes.

2026-07, v0.9: final editorial home

  • The home ending became a complete 03 / Important / 04 / colophon sequence with a separate ink safety panel.
  • A compact author block added his product and FinTech and AI Practice roles, 20 years of experience, and the scale of his work without using a photo.
  • The home kept its editorial cover while /method, /en, /about, and other pages kept a quieter reading mode. The new sections worked in both themes, on mobile, and in Site A, with no new network requests.

2026-07, v0.8.1: one shell

  • Home section 03 became a calm sage paper surface with a technical grid and cobalt glow.
  • The dark header became common to home and inner pages. Only the home kept the stronger editorial scene, while articles kept the readable v0.7 system.
  • Display heading line height increased so uppercase Russian diacritics no longer touched the line above on wide or mobile screens.

2026-07, v0.8: editorial stage

  • The home gained an ink and cobalt split hero, large type, a technical scale, and orange CTA tiles in real HTML text.
  • The principle became sage manifesto 01. Four entries and the route became dark editorial modules.
  • The wide layout gained a stronger cobalt frame, while mobile changed to a diagonal composition. Inner pages kept the quieter v0.7 field and their existing routes.

2026-07, v0.7: material field

  • A cold mineral field of off-white, sage, and muted cobalt appeared outside the book column.
  • Content moved onto its own paper surface. The hero blended cobalt into ink with a soft grid light.
  • Inner headings gained a light cobalt field. Prose, cards, tables, and every route stayed unchanged.

2026-07, v0.6: alternative design

  • The Practicum moved to a strict technical visual system: newsprint background, ink text, deep blue, and orange accent instead of the warm gradient.
  • Large light headings, monospace controls, flat panels, and a dark page bottom replaced the old style.
  • The table of contents, reading progress, full-screen mobile menu, dark theme, routes, bookmarks, and printed QR codes continued to work.

2026-07, v0.5: visual update

  • The design kept its indigo and amber character but used calmer surface layers and a better dark theme.
  • Russian headings moved to Manrope because the old display font did not support Cyrillic. Fonts began loading from the site itself.
  • Long pages gained an "On this page" outline and reading progress. Mobile navigation became full screen, and every page gained a footer. Routes did not change.

2026-07, v0.4: role strategies became action plans

  • All 13 role strategies gained a first-round start, work that stays human, review points, a stop threshold, and two next routes.
  • The final RU pass reduced unsupported numeric claims and connected the governance packet: scoring, human review, agent contract, evaluation set, shadow mode, and stop threshold.
  • Week Inventory became usable on mobile. Living pages aligned their verification dates, and build and route checks moved into CI.

2026-07, v0.3: progress layer

  • Maturity Route introduced five stages with an artifact and exit condition.
  • My Profile became local first. Browser tools could save progress, show the current stage and next step, export Markdown, and reset data stored only in the browser.
  • This changelog began recording Practicum versions.

2026-07, v0.2: from a library to a system

2026-06, v0.1: original release

  • Volume 1 and 2 Workbooks, 42 worksheets, 12 profession hubs using GDPval, the Anthropic Economic Index, and O*NET, nine browser tools, beginner materials, CLEAR prompts, dated living pages, and the One-Hour Book.

To suggest an improvement or send a case, see how to submit it. Back: home.

What is new in the Practicum