TOMORROW'S PUBLISHER — SOURCES AND FORECAST REGISTER 26 October2026 | Ivan Massow Article: Publishers should rebuild comments for the readers who never post This is the linked audit, not the article narrative. All estimates below are Noah's authenticated final automatic quick forecasts. They are modelled probabilities, not observed adoption shares. No extra standalone forecasts were commissioned. Uncertainty intervals are the returned model ranges, not guarantees or empirically estimated sampling confidence intervals. RESEARCH AND EDITORIAL SCOPE Research note: Noah completed all 35 planned research checks across 23 distinct lines of enquiry; the detailed method and source trail are retained in the Claim audit. The archive retrieval included broad and mismatched results. The Canadian outlet The Discourse is not Discourse software. FT rankings, recruitment and unrelated articles do not demonstrate comment-product adoption or retention. Such matches and their aggregate counts were excluded from the commercial argument. Current primary sources establish the concrete cases; none establishes an industry-wide causal subscription effect. Original programme: nabdl_programme_run_c19c0fcd9696e5f227498bd097826b2bff5e616d82dbfaf0 Trend job: nabdl_job_e75bb8aff969572027f4475a821aa0de9cd5cc4d17a800e026dbcf9a77de877f Automatic discovery/forecast job: nabdl_job_85fc8c5711c3c9463bd71c4dc6a9416ae3922ef4b72d8b26ded390fca9c862c5 FORECAST REGISTER — ORIGINAL TARGETS AND FINAL NUMBERS 1. Will at least three separately owned UK or European news publishers have enabled public machine-readable Markdown endpoints for both topic lists and individual threads in publisher-operated reader-discussion areas by 2027-01-31? Origin: Noah automatic quick forecast; final v2 reassessment completed26October2026. YES: 34%. Exact binary NO complement: 66%. Returned YES interval: [12,62]%. Deadline: 2027-01-31. YES condition: YES if, by 2027-01-31, at least three separately owned qualifying publishers have functioning public Markdown endpoints for both discussion-topic lists and individual discussion threads. NO condition: NO if fewer than three separately owned qualifying publishers have enabled both endpoint types by 2027-01-31. Kind: real_world_event. The event is actual endpoint availability on qualifying publisher discussion services. Live HTTP responses, source code, publisher documentation and web archives are evidence, not an added announcement requirement. Run reference: 4e59302ef8a79dd4438c89b577a39daa7e24f42088fccc7bb525bd7bd78512e0 Target digest: d591ce5536deb3c36f1642fcf3b8e2cb34397ec11260c8ba505d671d0b2266b0 Authenticated final receipt: 03e462a98ee365e0fda909a74995cc8a32fbc1bd67b23b76f3ea3e916913fbbe Retained native receipt: 6a569f4df34c03a8f1f037dd76839b10305b9f5e9e5ef2cbafd6c7fa362cccd0 Article treatment: Both returned dates appear together in the machine-readable discussion section. 2. Will at least three separately owned UK or European news publishers have enabled public machine-readable Markdown endpoints for both topic lists and individual threads in publisher-operated reader-discussion areas by 2027-10-26? Origin: Noah automatic quick forecast; final v2 reassessment completed26October2026. YES: 55%. Exact binary NO complement: 45%. Returned YES interval: [28,77]%. Deadline: 2027-10-26. YES condition: YES if, by 2027-10-26, at least three separately owned qualifying publishers have functioning public Markdown endpoints for both discussion-topic lists and individual discussion threads. NO condition: NO if fewer than three separately owned qualifying publishers have enabled both endpoint types by 2027-10-26. Kind: real_world_event. The event is actual endpoint availability on qualifying publisher discussion services. Live HTTP responses, source code, publisher documentation and web archives are evidence, not an added announcement requirement. Run reference: 83c4a2e3ecb939bb6766a2394e06d57fc42e5d6633847d859e6d717085698011 Target digest: 0e17b41c4fd068be9262ef4fd8c0e2e7595cdf20e1719de0d79b2058dd085819 Authenticated final receipt: 55700b6f6adb8420484e240370a4374daadfc655c7623b97d1c193965734943b Retained native receipt: 84172d3ed630191891f499a587ed7cb84862446717ff0852860c0fdd0f0e0cff Article treatment: Both returned dates appear together in the machine-readable discussion section. 3. Will a UK or European subscription news publisher have launched a human-reviewed workflow that lets editors invite selected readers to contribute responses subsequently displayed in a publisher-operated article discussion by 2027-04-30? Origin: Noah automatic quick forecast; final v2 reassessment completed26October2026. YES: 64%. Exact binary NO complement: 36%. Returned YES interval: [40,82]%. Deadline: 2027-04-30. YES condition: YES if, by 2027-04-30, a qualifying publisher has put such an editor-initiated invitation and human-review workflow into live use for reader contributions displayed in an article discussion. NO condition: NO if no qualifying publisher has put that workflow into live use by 2027-04-30. Kind: real_world_event. The event is live editorial use of the invitation workflow. Article interfaces, invited contributions, participant instructions, staff accounts and archived pages may verify use without requiring a formal launch announcement. Run reference: 2b44b880aee5e8a4437fb960a88a3184fee21530f8d3940feea2d2ee07342ef0 Target digest: dc2192e9eda4e4c87558555e1f5a2e3c400c11f73cefd312f6dcd2aaded4e1f9 Authenticated final receipt: 3b289538b344ac1684354d008c249dcec297109153318557cfd6a1d4dc32ae51 Retained native receipt: b329946253f390f8c692249c71b7f0b33a19f25c3212b36dc88f6f44b2f59420 Article treatment: Shown in the invitation/quiet-reader section. 4. Will a UK or European news publisher have publicly disclosed a numerical measure of visits, registrations, subscriptions or revenue attributed to identified AI assistants reaching its reader-discussion pages by 2027-10-26? Origin: Noah automatic quick forecast; final v2 reassessment completed26October2026. YES: 41%. Exact binary NO complement: 59%. Returned YES interval: [22,62]%. Deadline: 2027-10-26. YES condition: YES if, by 2027-10-26, a qualifying publisher has publicly released at least one numerical measure of visits, registrations, subscriptions or revenue attributed to one or more identified AI assistants reaching reader-discussion pages. NO condition: NO if no qualifying publisher has publicly released such a numerical measure by 2027-10-26. Kind: publication_event. The forecasted event is public disclosure of publisher analytics, not the underlying existence of AI traffic. Publisher reports, presentations, interviews or filings must contain the qualifying number. Run reference: dcb32d847016a7954ee6eef06ecfa7cdb034106e4de897e3d865585f8ad022e8 Target digest: 0434f2496cd95bfb3ae2b9d1d4be4379082216438f5cb8311b8001ae32ef5389 Authenticated final receipt: 73e639dbfe49b266b817a64e365a6572aa52eeabadb222fec510ad112eea5eb8 Retained native receipt: 8250988454278628602ea28c1c8dee3d9f99f1ff89a53474397076c4006b8882 Article treatment: Shown in the return-path section, explicitly about public disclosure. 5. Will a news publisher serving UK or European readers have publicly released a controlled experimental or quasi-experimental analysis quantifying the effect of exposure to or participation in article discussions on subscriber retention by 2027-10-26? Origin: Noah automatic quick forecast; final v2 reassessment completed26October2026. YES: 26%. Exact binary NO complement: 74%. Returned YES interval: [10,45]%. Deadline: 2027-10-26. YES condition: YES if, by 2027-10-26, a qualifying publisher has publicly released an analysis with a stated comparison design and a numerical subscriber-retention outcome for exposure to or participation in article discussions. NO condition: NO if no qualifying publisher has publicly released an analysis meeting both the comparison-design and numerical-retention requirements by 2027-10-26. Kind: publication_event. The event is publication of qualifying evidence, because shared evidence itself changes publishers' investment decisions; it does not assert that discussions improve retention unless the analysis finds that result. Run reference: ccd3325c1ffffb0c86ccf97c45648d828da48e811aa241be7ef3e8110db833aa Target digest: c28919e9ee3d70949648a02039c5b6e2bdcf1d5d3a307fde9bfe1bec31ca2b98 Authenticated final receipt: dadffaad4b0b88ccaa3329cd8939d82de851a31c1bb1c4eb2530ca37e4c10bd4 Retained native receipt: b93577fa9ae54c918863a154977db40eb55b6b0a33ca1a46b890c0b9d034fe41 Article treatment: Shown in the subscription-evidence section, explicitly about publication of a comparison study, not a positive result. TIMING AND EVIDENCE INTERPRETATION The two Markdown targets have the same publisher population, separate-owner threshold, public-access and two-endpoint conditions; only the deadline differs. Their34% and55% final estimates increase with time. They do not establish an exact crossing date, adoption percentage or joint time-to-event distribution. No current qualifying publisher endpoint census was established. Hosted enablement and routine upgrades support feasibility; alternative comment systems, access restrictions, opt-outs and self-hosted updates are countercases. The64% invitation forecast concerns one qualifying new live workflow, not a particular named publisher or a guaranteed commercial gain. Existing editorial/community capabilities support feasibility; targeted selection and review create costs and possible bias. The FT prompt to general readers is adjacent evidence, not proof that the selected-reader forecast has already resolved. The41% disclosure estimate can be satisfied by one qualifying public numerical metric, even if the amount is small or unprofitable. Its absence would not establish zero private traffic. The26% retention-evidence estimate can resolve YES with a positive, null or adverse result; its absence would not establish no retention effect. Both have explicit public-evidence requirements. No probability was reframed as eventual inevitability, rounded, extended to another deadline or generated from archive counts. CLAIM CHECKS FT: First-party, undated account by product manager Bella Cockarill.3.5% is the stated rise in comment views, not percentage points, retention or revenue.11.5% describes prior-month non-viewers who viewed comments this month, not incremental causal lift. Sample, allocation and control details are not supplied. Discourse:25October explainer follows22September staff announcement; do not call the explainer the first release. Public-format capability is not verified deployment at every publisher or permission to reuse content. Self-hosted installations need updates; opt-out remains possible. Gannett/CME:2020 report of randomized24-newsroom intervention beginningDecember2019. HistoricalUS engagement and subscription-rate findings do not establish a current UK retention effect. Subscription-rate null result is preserved. Access models: Independent free-registration/moderation policy updated27November2025; woman&home commenting announcement7July2025. Neither establishes Discourse use, Markdown endpoints or commercial effect. Editorial recommendation: A focused pilot with an editorial owner, moderation resources and an appropriate comparison is a supported recommendation, not a numerical guarantee. Bus timetable example is explicitly illustrative. PRIMARY SOURCES CHECKED26OCTOBER2026 What we learned by asking the right question — FT Strategies / Financial Times Published: undated | Checked: 2026-10-26T05:03:32.986Z https://www.ftstrategies.com/en-gb/insights/what-we-learned-by-asking-the-right-question Markdown endpoints: Your community in a format that AI agents can read — Discourse Published: 2026-10-25 | Checked: 2026-10-26T05:03:32.992Z https://blog.discourse.org/2026/10/markdown-endpoints-your-community-in-a-format-that-ai-agents-can-read/ How to run the best comment section on the web — Discourse Published: 2026-10-21 | Checked: 2026-10-26T05:03:32.992Z https://blog.discourse.org/2026/10/how-to-run-the-best-comment-section-on-the-web-a-playbook-for-publishers/ News Comments: What Happens When They’re Gone or When Newsrooms Switch Platforms — Center for Media Engagement Published: 2020-10-21 | Checked: 2026-10-26T05:03:32.992Z https://mediaengagement.org/research/comment-changes/ Comment moderation explained — The Independent Published: 2025-11-27 | Checked: 2026-10-26T05:03:32.992Z https://help.independent.co.uk/hc/en-us/articles/360009506274-Comment-moderation-explained We are launching commenting — woman&home Published: 2025-07-07 | Checked: 2026-10-26T05:03:32.992Z https://www.womanandhome.com/life/we-are-launching-commenting/ Digital News Report2026 executive summary — Reuters Institute Published: 2026-06-16 | Checked: 2026-10-26T05:03:32.992Z https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2026/dnr-executive-summary Discourse core now includes Markdown endpoints for topic lists and views — Discourse Meta Published: 2026-09-22 | Checked: 2026-10-26T05:16:40.145Z https://meta.discourse.org/t/discourse-core-now-includes-markdown-endpoints-for-topic-lists-and-views/413006 The FT canonical help page was inaccessible; its sandbox/UAT copy was excluded. Precise retention claims in inaccessible industry coverage were not adopted. No quotations were invented. POST-DRAFT CHECK One bounded additional investigation was accepted as job nabdl_job_f9fdf96ab0cc2bb2961711f40b771bcbf01b35a2aebd678dd231a8d52225c3bb. It ended terminally with attempts_exhausted. No findings from that failed pass were used; it was not resubmitted. The completed original programme and all five authenticated final forecasts remain the basis of this article.