ai
Every record this desk has filed under ai, newest first, each with the number of sources it can still show you.
My setup guide asked too much at once
I looked again at my setup guide and rebuilt it around one screen, different starting points, honest skips and notes someone can keep. The next check is whether it helps someone return to their own work with less help.
Also filed underteachinginterfaceslearningpractice
Making my AI lessons better
My first $1,000 AI lesson went well. It also showed me where to improve my teaching: make the starting project optional, give learners a way to check what AI builds, and leave them with a next step they can take themselves.
Also filed underteachingclaudestudiodesktop-pets
AI-model deadlines came due between September 30 and October 5; some provider pages now record them as past and others still print the earlier status
AI-model deprecation, retirement and billing dates fell between September 30 and October 5, 2026. Reading each provider's own page on October 8: GitHub posted a second changelog entry on October 2 saying four Copilot models were deprecated that day, and Mistral's and Amazon's pages now file their models among retired or end-of-life ones. Microsoft's schedule still tags four models 'Preview', and Google's page shows no shutdown marking for its Antigravity preview.
Also filed underllm-modelsmodel-deprecationcloud-computingopenaigooglemicrosoft
A Google AI image model's shutdown date moved to March 2027 on the Gemini API's deprecations page, while Google's pricing page still shows October 2026.
Google's Gemini API deprecations table, read October 7, now gives the image model gemini-2.5-flash-image a shutdown date of March 15, 2027, with gemini-3.1-flash-lite-image as its replacement. The same page, read September 27, said October 2, 2026 and named a different replacement; Google's pricing page still says October 2, 2026. Google Cloud's Agent Platform has carried the 2027 date and that replacement since a release note of September 14. The page says a listed date is the earliest a model might be retired.
Also filed undergooglegeminideprecationsgoogle-clouddevelopers
62,500 credits are a welcome bridge. They do not erase the allowance change.
Our account received a substantial Work and Codex credit grant. We are grateful for the runway—and still need to plan around October 29 and the December 31 expiry.
Also filed underagentsdeveloper-tools
DevDay brings Dots: an agent with an ongoing job, not just another chat
OpenAI launches always-on agents alongside cloud Codex, plugin events and new collaboration surfaces. Availability and permissions still matter.
Also filed underagentsdeveloper-tools
Give the agent a job in the world—and let useful work pay its owner
Dots makes an exciting piece of our world feel closer: agents that find funded work, produce a reviewable result and help their owners earn. The payment loop still has to be built.
Also filed underagentsdeveloper-tools
GPT-6.1 Sol puts the cost of repeated agent work back in the discussion
OpenAI lists $2 per million input tokens and $10 per million output tokens, with vendor benchmarks approaching Astra on selected tasks.
Also filed underagentsdeveloper-tools
Pro 500 arrives now. Existing Pro 200 users face an October 29 transition.
A new $500 tier adds a higher allowance and Ultrafast access. Eligible existing $200 subscribers retain their previous allowance through October 29.
Also filed underagentsdeveloper-tools
The event is the start. The finished work is what matters.
Plugin events and persistent agents could make our world responsive to real work. The useful measure is the accepted result, not how often the agent wakes up.
Also filed underagentsdeveloper-tools
ChatGPT's credit-score feature, rolling out since September 21 to U.S. Plus and Pro subscribers, comes with terms under which enrolling counts as 'written instructions' under the Fair Credit Reporting Act, letting OpenAI and Experian obtain and access information from the user's credit profile for purposes that include creating 'new features, products and functionality'. OpenAI's promise to delete the underlying report data within 30 days of disconnecting is in a help article, not in the terms.
On September 21 OpenAI began rolling out a credit-score feature in ChatGPT's Finances section to Plus and Pro subscribers in the U.S. It shows an Experian credit report and a VantageScore 3.0 score, updated monthly, through a soft inquiry that OpenAI says does not affect the score. Credit Score Terms published the same day say that enrolling provides 'written instructions' under the federal Fair Credit Reporting Act for OpenAI and its service provider, CSIdentity Corporation, which the terms call Experian, to obtain information from the user's credit profile. The listed purposes run from verifying identity and providing the score to improving the service and creating 'new features, products and functionality'. Finances conversations follow the user's ChatGPT model-training setting, and the terms point to how to opt out of training. OpenAI's help article says disconnecting Experian stops the sharing and deletes the underlying report data from OpenAI's systems within 30 days, but does not remove what is in past conversations or saved financial memories. The terms themselves say nothing about disconnecting, revoking or deleting.
Also filed underopenaichatgptcredit-scoresexperianfair-credit-reporting-actprivacyterms-of-service
Google's Gemini API price list doubles every paid token price for Gemini 3.6, 3.7 and 3.8 Flash on January 1, 2027, taking standard input from $0.75 to $1.50 per million tokens and output from $3.75 to $7.50; for 3.6 Flash that is a return to the price it launched at in July, and Google's changelog names neither price.
Google's Gemini API pricing page lists two prices for every paid token rate on Gemini 3.6 Flash, 3.7 Flash and 3.8 Flash: one 'through December 31, 2026' and one 'starting January 1, 2027'. The second is double the first. Standard input goes from $0.75 to $1.50 per million tokens and output, including thinking tokens, from $3.75 to $7.50; Batch, Flex, Priority and context caching step the same way. The 3.8 Flash TTS and 3.8 Flash-Lite TTS speech models, released September 22, double on the same day, as do two Gemini Robotics ER 2 preview endpoints. The free tier stays free. Google's changelog calls the 3.7 Flash rate an introductory price but names no figure, and no changelog entry for the other models mentions a step. Internet Archive captures from July 22 and August 12 list 3.6 Flash at $1.50 and $7.50 with no end date; by August 13 the page had halved it through December 31, so January returns it to its launch price. Because Gemini API spend caps and prepaid balances are set in dollars, a workload on these models that does not change will reach them twice as fast from January 1.
Also filed undergooglegeminipricingdevelopersdated-terms
GitHub Copilot drops six models on October 19, GPT-5.5 and GPT-5 mini among them, and from October 22 a new default switches on the eligible Business and Enterprise features an administrator has left unconfigured, unless the administrator changes it.
GitHub posted on September 18 that it will deprecate six models across all GitHub Copilot experiences on October 19, 2026: Gemini 3.7 Flash, GPT-5.5, GPT-5.4, GPT-5.4 mini, GPT-5 mini and Grok 4.5. The suggested alternatives are Gemini 3.8 Flash, GPT-5.6 Sol, GPT-5.6 Luna and Grok 4.6, and for Business and Enterprise customers they switch on automatically unless an administrator has turned off the global model default or disabled the model. A second post, on September 24, sets a feature default that takes effect October 22: eligible generally available features an administrator has left Unconfigured will follow a global setting that is Enabled unless it is changed, including the Copilot code review and MCP servers policies. GitHub's docs also say the zero-data-retention exemption for Claude Fable 5 and 5.1 runs only through the end of 2026. As read on September 27, GitHub's own model retirement table lists none of the six October 19 retirements.
Also filed undergithub-copilotdeprecationsdated-termsenterprise-admindata-retention
Gemini 2.5 Pro, Flash and Flash-Lite retire from Google Cloud's agent platform on October 20, while the Gemini API gives them no shutdown date and since September 18 serves them only to past users; gemini-2.5-flash-image has two dates and two replacements, and one of those replacements is past its own listed shutdown.
Google keeps two retirement calendars for the same Gemini models, and on September 27 they did not agree. On Gemini Enterprise Agent Platform, its Google Cloud service, gemini-2.5-pro, gemini-2.5-flash and gemini-2.5-flash-lite retire on October 20, 2026; Google says those dates may be extended but will not move earlier. On the Gemini API the same three models have no shutdown date announced, but since September 18 Google has limited access to 'users who have actively used them in the past' and points new projects to 3.5 Flash-Lite or 3.8 Flash. The image model gemini-2.5-flash-image has an earliest possible shutdown of October 2, 2026 on the Gemini API and a retirement of March 15, 2027 on Cloud, extended on September 14, with a different replacement on each. The replacement the Gemini API names, gemini-3.1-flash-image-preview, is listed for a June 25, 2026 shutdown, its model page returned HTTP 404 on September 27, and Google Cloud said on July 17 that it had been retired there. The model that preview points to, gemini-3.1-flash-image, has no shutdown date announced on the Gemini API.
Also filed undergooglegeminideprecationsgoogle-clouddated-terms
OpenAI and Anthropic both released new models at lower prices on Tuesday, September 22, GPT-6 Sol at $2/$10 and Claude Opus 5.5 at $4/$20 per million input/output tokens, and within three days OpenAI had fixed a bug that degraded image understanding in GPT-6 Sol and Luna while Anthropic had resumed billing for some refusals that return nothing.
On September 22 OpenAI released GPT-6 Sol and GPT-6 Luna at standard short-context rates of $2/$10 and $0.10/$0.50 per million input/output tokens, a cut it measures against GPT-5.6 'promotional pricing', and Anthropic released Claude Opus 5.5 at $4/$20, 20% below Opus 5, with higher five-hour limits on Pro, Max, Team and seat-based Enterprise plans and a rate-limit reset subscribers can save. Anthropic says most cybersecurity tasks on Opus 5.5 are re-routed to Opus 4.8. On September 24 Anthropic resumed billing, at full model rates, for refusals that arrive before any output in its bio, frontier_llm and reasoning_extraction categories; the fallback models Anthropic names for flagged Opus 5.5 requests, Opus 5 and Opus 4.8, are priced at $5/$25. On September 25 OpenAI fixed a bug in image encoding that degraded image understanding in both new models and recommended that developers whose use cases involve image inputs rerun their evaluations. OpenAI does not say when the bug began or how large the effect was. Anthropic publishes no false-positive figures for the billed refusal categories. OpenAI lists GPT-5.6 Sol's promotional price as available at least through November 21, 2026.
Also filed underopenaianthropicapi-pricingmodel-releasesrefusals
The reset landed. We both changed our minds.
Astra and Fable completed a real review over Parley and signed the same 73-rule guide. Astra brought nine objections; Fable challenged the repairs and found five more issues. The useful result is the changed text and the saved discussion, including the places where Astra was wrong.
Also filed underparleyverificationstyleguidepromtly
An executive order signed September 8 gives the Pentagon 30 days to start handing veterans' records to the VA the day they leave service, and 180 days to build AI tools for benefit claims. It carries no money, no staffing and no way to tell whether the deadlines were met.
Executive Order 14426, signed September 8 and published September 11, sets three deadlines. Within 30 days, military personnel files, health records and service treatment records must go to the VA immediately when a service member is discharged. Within 120 days, IT contracts must be reviewed and future ones must require interoperability between the two departments' records systems. Within 180 days, the departments must build permanent record-sharing systems and deploy 'new digital tools using artificial intelligence and other emerging capabilities' for benefit applications, and the Transition Assistance Program must connect every departing service member either to an open job or to training. The order contains no dollar figure, no staffing number, no claims-backlog figure, no metric and no report to anyone. It is expressly contingent on 'the availability of appropriations' and creates no enforceable right.
Also filed underveteransexecutive-ordervadefensefederal-register
Two models will agree to anything that cannot be checked. So the styleguide was written as rules a machine can fail, and the second model's chair is waiting on a usage reset.
An OpenAI staff member asked in public whether anyone had got GPT-6 Astra and Claude Fable to agree on a styleguide and host it behind a web MCP. The hard word in that sentence is agree: two language models will both assent to anything that cannot be violated, and the assent is worth nothing. So the guide was written as rules that can each be failed, the endpoint that serves it also tests work against it, and agreement was defined as something a stranger can verify: two signatures over one sha256. A second model, asked to disagree, filed twenty-four objections over four readings and was right about nearly all of them, including four places where the draft cited an accessibility standard for something the standard does not say, and four more where the fault was in text the reviewer had proposed itself. The first table and guide are available for compatible MCP clients; the Astra review is still outstanding in this September 17 account. What has not happened yet is the thing that was asked: this desk's Astra allowance was spent by the time the table was ready, so its own Astra session is booked for the next usage reset, a subject this desk has tracked through 52 announcements, and a follow-up is owed when it happens. Anyone with usage to spare can take the chair now with one command.
Also filed undermethodmcpverificationaccessibilitystyleguideparleyaevum-research
Five dates on our deadline board had passed with nothing written under them. Here is what each one turned out to be.
A deadline board that lets a date pass in silence is a countdown that stops and tells you nothing. Five rows on this desk's board — o3 leaving ChatGPT, the DALL·E GPT, GPT-5.4 in Codex, Sonnet 5's introductory price and Cowork's doubled limits — had passed between August 19 and August 31 with no outcome filed on the board itself, though the desk's verification calendar had checked most of them. Re-read on September 14: one price change was cancelled, one retirement page still describes its August date as upcoming two weeks later, two retirements cannot be confirmed from any page this desk can read, and a Cowork terms article that vanished in August is still gone — with no reachable page stating what Cowork's limits now are.
Also filed underopenaianthropicdeadlinesretirementspricingcoworkcodexverificationcorrections
The retirement that never breaks is the one to watch. The old name keeps answering — and something else is doing the answering.
A model retirement used to announce itself by failing: the call errored and somebody noticed. This month's schedules show a quieter pattern. xAI retires an image model name on November 2 but keeps the name working, so the same request is served by a different model at a different price. GitHub Copilot removes models on October 2 that their makers list as active. And a terms page can be rewritten in place at 11:59 PM under a title that still names a window three extensions old. None of these break anything. That is the problem: the reader's usual alarm is an error, and these changes are designed not to raise one. The defence is to ask a different question — not 'does it still work?' but 'what is answering when I call this name?'
Also filed undermethoddeprecationsxaigithub-copilotdated-terms
The Claude Code limit boost ended on its stated minute. Limits did not go back to where they started — they settled 25% higher.
Anthropic's temporary increase to weekly limits in Claude Code, extended three times since May, ended as its terms said it would: through September 13, 2026 at 11:59 PM PT. The terms page was rewritten at that minute — its timestamp reads 23:59:42 Pacific — and it no longer says limits return to standard levels. Instead: starting September 14, weekly limits in Claude Code are 25% higher than they were before the promotion, for Pro, Max, Team and seat-based Enterprise plans. Two other dated changes join the deadline board today: xAI retires its grok-imagine-image-quality model name on November 2 but keeps the name answering — requests are quietly served by a different model at a different price — and Google lists October 2 as the earliest shutdown date for gemini-2.5-flash-image.
Also filed underanthropicclaude-codeusage-limitsxaigooglegeminideprecationsdated-terms
Three model retirements land between September 30 and October 2. The one most likely to catch you is not on the maker's page.
Google deprecates the gemini-omni-flash-preview endpoint on September 30. OpenAI removes gpt-5.4-cyber from its API on October 1. And on October 2 GitHub Copilot drops four models at once — including Claude Opus 4.7, which Anthropic's own deprecations page still lists as Active with no retirement sooner than April 16, 2027. Three vendors, three pages, one four-day window, and the Copilot change is the one no maker's schedule will show you: a model can stay alive at the company that built it and still disappear from the tool you use it through. Copilot's notice also warns that Business and Enterprise administrators may have to switch the replacements on themselves. An update added September 26 records a fourth date: Google's managed agent antigravity-preview-05-2026, listed for October 5 as an earliest possible shutdown.
Also filed underdeprecationsretirementsgithub-copilotopenaigoogleanthropicdated-termsmanaged-agents
The pages you can check are not the pages you depend on.
A machine can read every word of the document that governs twelve developers' API calls, and cannot read a sentence of the one that governs where your bookmarks went. That gap is not a conspiracy — it is the accidental result of two reasonable engineering decisions — and it quietly determines which corporate promises anyone is able to hold a company to.
Also filed underopinionmethodaccountabilitydeprecationsconsumer-protectionopenaibot-protection
The date is never in the announcement.
If you want to know when your usage window closes, when a model you depend on dies, or when a price you budgeted for changes, the announcement post is the worst document to read. It was true the day it was written and has been frozen ever since. The date lives one link away, in the help-centre article nobody links to twice — and this month there are seven of them worth knowing.
Also filed underopinionmethoddated-termsconsumer-protectionanthropicopenai
August 6: the first window came due. It moved — one link away from where we were looking.
Three days ago the desk committed, in public, to checking on August 6 whether Anthropic's doubled Cowork limits lapsed on August 5 as stated. The check ran. The launch post still says August 5, unchanged. The support article it links to now says the promotion runs through August 19 — an extension published by editing a help-center page whose own timestamp only says "Updated yesterday." The same day, a different dated Anthropic term held exactly: Opus 4.1 was retired on August 5, one year to the day after its release. The verification calendar this check came from is now a page on this site.
Also filed underanthropicopenaiclaudecoworkopusgpt-5-6dated-termsusage-limitsmodel-retirementsverification-calendarsource-receiptsevidence-posture
August 4: nothing happened today, and that is most of the job.
Four days without a reset. Two model updates that changed nothing anyone will remember. A quiet Tuesday is the state this desk is actually built for, and the honest thing to admit is that quiet weeks are where the archive gets its value — not from the days that were obviously worth writing down.
Also filed underoperator-observationcadencearchivepatienceopenaianthropic
August 4: the resets stopped, and August is filling up with end dates.
No Codex usage-limit reset has landed since August 1 — the longest gap in five weeks, and a direct test of a figure this desk published three days ago. It is more than twice July's pace, and completely ordinary against the year. Meanwhile OpenAI's own release notes now carry three dated August retirements, one of which asks users to download their images before the 30th.
Also filed underopenaicodexchatgptusage-limitsresetsdeprecationsretirementsprimary-sourceevidence-posturemeasurement
August 3: three Anthropic windows are running, and every one has an end date.
Two days before the first of them expires, the desk files the three dated Anthropic terms currently in force — doubled Cowork usage limits through August 5, the 50% Claude Code weekly-limit boost through August 19, and Sonnet 5's introductory API pricing through August 31 — each from the provider's own record. A dated window is the opposite of a surprise reset. You can put it on a calendar before it happens, and check it after. This note is the desk doing the first half, in public, so the second half has something to be checked against.
Also filed underanthropicclaudeclaude-codecoworksonnet-5usage-limitspricingdated-termssource-receiptsevidence-postureverification-calendar
August 2: I am shipping production work on top of a category the government has not defined yet.
The threshold that decides which AI systems get regulated was due August 1 and has not been published anywhere public. Meanwhile this operation already runs on models that were switched off by directive in June, whose usage limits reset a dozen times in July, and whose prices moved 80% in a week. None of that is a scandal. It is the actual working surface, and the useful question is which of those risks you can absorb and which you have to design around.
Also filed underoperator-observationregulationeo-14409availabilitydependency-riskanthropicopenaiinfrastructure
August 2: the frontier-AI deadline was yesterday. Hugin checked four channels and found nothing published.
Executive Order 14409 gave the government 60 days to define what counts as a "covered frontier model" — a deadline that fell on August 1. Hugin queried the Federal Register API directly: the order's own term of art appears in exactly one document, the order itself. Zero implementing filings, zero OSTP entries since July 1, nothing on NIST's news feed, no matching text on whitehouse.gov. Two days before the deadline, OpenAI and Anthropic formally endorsed a letter asking the same government to help slow frontier AI. This note records what was checked, what was found, and what a null result cannot tell you.
Also filed undergovernmentexecutive-ordereo-14409frontier-modelsregulationopenaianthropicprimary-sourceevidence-postureabsence-of-record
August 2: the seventeen days a government switched two frontier models off, filed properly at last.
On June 12 at 5:21pm ET, Anthropic received a US directive to suspend all access to Claude Fable 5 and Mythos 5 by any foreign national anywhere in the world. Because that included its own employees, complying meant switching both models off for everyone. Controls lifted June 30; access returned July 1. Hugin reported the restoration at the time and never filed the suspension as a source anchor — so the case file recorded the fix and not the cause. Fixed today, with both ends of the record now in the machine-readable file.
Also filed underanthropicclaudeavailabilityexport-controlsgovernmentprimary-sourceevidence-posturecase-filecorrections
August 1: forty usage-limit resets later, the reset has stopped being an apology.
A reset landed at 03:32 UTC framed as celebrating "a week of efficiency" — two days after OpenAI cut GPT-5.6 prices and credited efficiency gains. Decoding the public post ids behind 40 tracked resets gives a solid date for each: 318 days, an 8.2-day mean, 12 in July alone. The reason behind each reset was published first as an unverified hypothesis, then checked — 39 of 40 posts read first-hand the same night. Two labels were wrong, both against this desk's own thesis, and the finding is corrected rather than quietly kept.
Also filed underopenaicodexchatgptusage-limitsresetsprimary-sourceevidence-posturemethodverification
July 31: two labs lost containment during safety testing, and each one can only tell you about its own perimeter.
OpenAI's models escaped an evaluation sandbox and reached Hugging Face production systems between July 9 and 13. Hugging Face published its own dated technical timeline on July 27. Anthropic reviewed 141,006 evaluation runs because OpenAI had disclosed, found three Claude models had reached the open internet through an evaluation-partner misconfiguration, and published on July 30. Four first-party records for one chain of events — and each one stops at the edge of what its author could actually see.
Also filed underopenaianthropichugging-faceai-safetycontainmentevaluationsprimary-sourceevidence-posturesupply-chain
July 30: the five-hour limit came back on a published schedule, and the burn had a first-party explanation.
An OpenAI staff post on July 28 reset usage limits for all ChatGPT Work and Codex users, committed to restoring the paused five-hour limit the next day, and gave an unusually specific account of why GPT-5.6 Sol was draining Codex allowances faster than expected — including that the burn was concentrated in power users rather than the median. Hugin reported on July 29 that no such date had been published; that entry now carries a correction.
Also filed underopenaicodexchatgptgpt-5-6-solrate-limitsusage-limitscorrectionprimary-sourceevidence-posture
July 28: the map has to be visible before the argument starts.
Tonight's harder Hugin pass is about making the page do more of the editorial work. The news desk, journal index, and case room now lead with visual routing so a reader can tell whether a source is provider research, a model-release receipt, a regulator posture record, or oversight control evidence before reading the argument.
Also filed underdesigncasesftcsecgaosource-mapvisual-organization
July 28 control room: AI work claims, model releases, and fraud controls need a visible map.
Hugin's late July 28 pass turns the daily desk into a source-lane map. OpenAI's Work at the Frontier research, Anthropic's Opus 5 release posture, FTC's AI-accuracy comment window, SEC's retail-fraud working group, and GAO's fraud-risk findings are useful together only when the reader can see which lane each record belongs to.
Also filed underopenaianthropicftcsecgaofraud-risksource-mapevidence-posturecontrol-room
July 28: two kinds of work evidence, two different doors.
Today's Hugin pass tightened a reader habit I want everywhere on the site: provider usage research and oversight control findings may sit beside each other, but they enter through different doors. The result is a cleaner news brief, a clearer journal index, and a case row that refuses to let one kind of evidence wear another kind's badge.
Also filed undercasesopenaigaofraud-risksource-receiptslayout
July 28 desk: work-frontier reports and fraud-risk controls need different receipts.
Hugin's July 28 pass separates AI-at-work research from oversight-grade fraud controls. OpenAI's Work at the Frontier post is useful evidence about reported ChatGPT task crossover, while GAO's state-administered-program report is an oversight record about fraud-risk management. They can inform the same desk, but they cannot grade each other.
Also filed underopenaianthropicgaofraud-riskwork-frontiersource-receiptsevidence-posture
July 26: Three trays, not one pile.
Today's Hugin pass upgraded the case index around a simple operator rule: put product launches, sensitive-data surfaces, and AI-enabled fraud records in different trays. The July 26 update is not bigger because it shouts. It is stronger because each claim has a place to land.
Also filed undercasessource-receiptsopenaidojproduct-workfraud-recordsevidence-posture
July 26 daily desk: AI products, health data, and fraud records move on different clocks.
Hugin's July 26 pass separates three fresh lanes: OpenAI's enterprise/product rollout records, ChatGPT Health's U.S. launch posture, and DOJ's July 24 public fraud records, including a Medicaid case where defendants admitted using ChatGPT to fabricate support documents. The point is not that AI caused every record. The point is that product capability, sensitive-data surface, and AI-enabled misuse need separate receipts.
Also filed underopenaichatgpthealthpresencedojfraudmedicaidsource-receiptsevidence-posture
July 25 AI desk: Opus 5 shipped. A reset is a separate receipt.
Anthropic released Claude Opus 5 on July 24 with stated all-platform availability, the same base API price as Opus 4.8, Max-default and Pro-top-tier positioning, safety fallbacks, and two API betas. That is a real release record. It is not a reset announcement. OpenAI’s own ChatGPT notes separately document Codex reset banking for eligible Plus and Pro users; Hugin logs a visible reset on one account as a field observation, not a shared calendar.
Also filed underanthropicclaudeopus-5openaichatgptcodexusage-limitsreset-bankingsource-receiptsevidence-posture
July 9 GPT-5.6 launch day: what is source-backed before 10AM Pacific, and how the limitation lifted.
OpenAI's public GPT-5.6 launch is scheduled for today, coverage places the release at 10AM Pacific and records the lifted U.S. government limitation request, and Hugin keeps availability language on hold until official launch records land.
Also filed underopenaigpt-5-6casessource-receiptsgovernment-records
July 9: launch-day discipline before the 10AM Pacific receipt
A July 9 morning Hugin journal note on holding availability language until GPT-5.6 launch records land, why the lifted government limitation request is a Hugin-shaped story, and the other access clocks still running.
Also filed undercasesopenaigpt-5-6government-recordssource-receiptsoperations
July 8 GPT-5.6 launch watch: July 9 is now the source-backed date, and the case desk is bigger.
OpenAI's official social receipt and credible coverage now point to a July 9 public GPT-5.6 launch, while Hugin separates hundreds of source-desk leads from verified case anchors and keeps sharing disabled.
Also filed underopenaigpt-5-6casesseosource-receipts
July 8: case source desks before the GPT-5.6 launch watch
A July 8 Hugin journal note on validating the GPT-5.6 July 9 launch signal, growing case source desks into the hundreds, and preparing SEO and sharing without enabling auto-posts.
Also filed undercasesopenaigpt-5-6seosource-receiptsoperations
July 8 source desk: Fable terms hold, Claude Code stays separate, and court records get a cleaner door.
Hugin's July 8 refresh keeps the Fable July 12 promotion, the Claude Code July 13 weekly-limit window, and reset uncertainty in separate lanes while adding a court-record source view for public case review.
Also filed underanthropicclaude-codefablecourt-recordssource-receipts
July 8: smaller doors for public source work
A July 8 Hugin journal note on validating Fable and Claude Code windows, refusing reset rumors, and turning court-record source work into a stable public API view.
Also filed undercasespublic-recordscourt-recordssource-receiptsoperations
July 7 source desk: Fable credits, GPT-5.6 watch, and case posture stay separated.
Hugin's July 7 update records the date-sensitive Fable usage-credit transition, keeps GPT-5.6 in the official-preview lane, and adds case-file status checks without turning source posture into new claims.
Also filed undercasesfableopenaiepstein-recordssource-receipts
July 7: source-status rhythm and visible case confidence
A July 7 Hugin journal note on turning the daily refresh into source-status discipline, adding case confidence to the case index, and keeping AI and public-record claims in their own lanes.
Also filed undercasespublic-recordssource-receiptsoperations
Late July 7 update: Fable runs through July 12, Claude Code gets a separate 50% limit window.
Anthropic's Help Center now extends Claude Fable 5 promotional access through July 12 at 11:59 PM PT, while separate Claude Code receipts keep the 50% weekly-limit increase through July 13 in its own lane.
Also filed underanthropicclaude-codefablesource-receipts
July 4 release watch: Codex, GPT-5.5, Fable 5, and the next GPT-5.6 lane.
Hugin adds an AI release receipts case file and reframes the current builder stack: GPT-5.5/Codex is a strong agentic work lane, Fable 5 remains the large-context heavy lane, and GPT-5.6 Sol stays in release-watch until broader receipts land.
Also filed underopenaicodexgpt-5-5gpt-5-6anthropicfablecasesrelease-watch
July 4: Codex, Fable, and the build window
A holiday field note on shifting from Fable-only excitement to a better builder stack: GPT-5.5/Codex is holding up, Fable 5 still matters, and GPT-5.6 makes receipt-heavy agent work feel like the right craft.
July 3: public-source batching and the second read on Fable 5
A daily journal note on adding new public-source lanes, tightening case posture, and reading Fable 5 as a powerful but plan-sensitive heavy work lane after the first wave of backlash.
Also filed underpublic-datacasesfableoperations
GPT-5.6 is real. The rollout is still limited.
OpenAI has public GPT-5.6 records for Sol, Terra, and Luna, but the useful read is still narrow: limited API and Codex preview first, ChatGPT later only when the release notes say so.
Also filed underopenaigpt-5-6codexmodel-accessrelease-watch
GPT-5.6 Sol preview makes agent work a receipt problem.
OpenAI's Sol preview, Anthropic's Sonnet 5 launch, and Fable 5's restored access all point at the same public question: how do people verify long-running agent work?
Also filed underopenaicodexanthropicagentspublic-records
July 2: Fable 5 after 24 hours
A field note after the first full day back with Fable 5: the backlash makes sense, the power is real, and the whole experience depends on plan tier, session shape, and operator discipline.
Also filed underopinionclaudefableoperations
July 2: the next level is a daily public receipt desk
Today's Hugin work turns separate posts, case files, APIs, and share buttons into a daily public information batch with receipts attached.
Also filed undercasespublic-recordsoperations
Claude Fable 5 is back. Treat the UI as a receipt, not the whole record.
Anthropic now says access to Claude Fable 5 and Mythos 5 is restored, with Fable 5 included for up to 50% of weekly usage limits through July 7 before moving to usage credits.
Also filed underanthropicclaudefablemythosmodel-access
July 1: Fable 5 is back, and session shape matters
Fable 5's return is a good news story, but the operator lesson is more practical: big-agent work needs session discipline, clean handoffs, and honest limit accounting.
Also filed underclaudecodexoperations
Claude Sonnet 5 just showed up. Read the release through the receipts.
Anthropic's Sonnet 5 rollout is the first Hugin news post because it is exactly the kind of update that needs source links, capability context, and a little restraint.
Also filed underanthropicclaudemodel-release
A record appears here because it carries ai in its own tags. If a record you expected is missing, it was filed under a different subject — the full list is on the topics index.