Skip to content
Hugin

The archive

Search the record.

Reporting, commentary and the evidence behind the story.

Published records

1–24 of 37 matches

  1. Commentary

    A $70,000 fee gets a 30-day comment period; the paperwork that would record it gets 60, in one DATES line

    …would record the payment. The document calls the shorter window a "30-day public comment period" in its own text. November 9 is thirty-two days after publication, which is that same window landed on a…

    Read record summary

    DHS's Optional Practical Training fee proposal sets two deadlines in a single DATES line: November 9, 2026 for the rule, December 7 for its information collection. The rule's one sentence about those 30 days measures them against the time an F-1 student needs to change status or depart — notice, rather than time to comment. The Paperwork Reduction Act's 60-day notice provision carries an exception for a collection inside a proposed rule, so on the statutes read here, neither number is a floor someone set for the agency.

    Matched in title, summary, article text

  2. Commentary

    Nobody hid the black lung rule. On its last day, five comments had been filed. That gap is what everything we are building is for.

    Today is the last day to tell the Labor Department what you think of a rule about black lung. The rule decides how much money a coal company that insures itself has to lock away, so that the miners it employed…

    Read record summary

    Today is the last day to comment on a Labor Department rule deciding how much coal companies that insure themselves must set aside so miners with black lung still get paid if the company fails. The last time that money ran short, three bankruptcies left an estimated $865 million with a public trust fund. The rule has been public for sixty days. By Monday morning, five comments had been filed. The problem was never secrecy; it was delivery. This is a plain-English tour of how a small studio is building against that gap: a public record that carries its receipts, two AI models that only agree when they sign the same text, the shared plumbing that ties them together, and Cairn, a membership built to be audited that is close to welcoming its first members. Getting in will take a test. The goal it is setting out loud: ten million jobs in ten years, counted in public. Also: what is not finished, and what nothing measures yet.

    Matched in title, summary, article text

  3. Commentary

    An extension has no launch day, so nobody hears about it.

    …it costs somebody something real. The asymmetry A launch has a launch day. There is a post, a headline, a thread, a moment where the company gets to say look what we made . The machinery of telling people…

    Read record summary

    Anthropic gave Claude Code users twelve extra days of raised weekly limits and told nobody — not because it is hiding anything, but because an extension has no marketing moment. The people who lose from that are the careful ones: whoever actually read the terms and paced themselves toward a date that had already moved. Good news travels worse than bad news, and terms pages have no history at all.

    Matched in title, article text

  4. Commentary

    August 4, later: I was wrong about the quiet day.

    …whole of 2024 absent, in the most exhaustively documented public-records matter of the last decade. And it had a tell I should have caught by reading, long before any tool existed: the file recorded the…

    Read record summary

    Earlier tonight I wrote that nothing happened today and that quiet days are the archive's raw material. Then I pointed a new tool at the case files and it told me ten files were clean. There are eleven. The one it skipped was missing twenty months, and I had spent the evening arguing that the value of this desk is that it keeps the record.

    Matched in title, article text

  5. Commentary

    The most expensive thing in America is an hour to think

    …off into a picture. By morning most of it is gone, and I spend the next day trying to remember what I had already worked out. That stretch is not wasted time. In a 2021 study, people given a 20-minute rest…

    Read record summary

    This week I described a state of mind in my own words, then found William James had named two of the same marks in 1902. That points at a pattern in how brains work, and no machine can catch it yet, even with an implant. Meanwhile the hour people would think in is being priced out: by my arithmetic from federal data, day care up 166 percent since 2000 and tuition up 208, against pay up 137. Cash alone mostly bought rest. Here is what I would change, and what I am building.

    Matched in summary, source, article text

  6. Commentary

    July 1: Fable 5 is back, and session shape matters

    …deep work lane active, give it a real objective, and let the rest of the day orbit around that lane instead of competing with it. A practical loop has started to emerge: Start a fresh chat and describe the…

    Read record summary

    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.

    Matched in title, article text

  7. Commentary

    AEP is here, and I finally have more time for clients

    …it. This is my account of what we are building and how it is changing my day. More room for the conversation Most of my day-to-day work has been automated now, apart from talking to clients. That is a big…

    Read record summary

    Medicare's annual enrollment period starts today. This week at SilverPath, we pulled more than 350,000 doctor records and millions of plan connections for our quoting tool. Most of my routine work is now automated. The part I want to spend my day doing is still the conversation with a client.

    Matched in summary, article text

  8. Commentary

    The bank gets 90 days and a sentence that says it admits nothing. The immigrant gets 15 business days and a postmark.

    Two federal enforcement documents landed a day apart this week, and if you read them back to back you can see the whole machine. On October 8 the Office of the Comptroller of the Currency found that American…

    Read record summary

    In one week the federal government wrote down two theories of accountability. A bank its regulator found had processed about $13 billion in suspected trade-based money laundering activity from 2014 to 2025 got a negotiated $350 million penalty, 90 days to write a plan and the right to neither admit nor deny. A person who willfully fails to leave after a removal order faces up to $998 a day, assessed at the maximum as a uniform practice, with 15 business days to get an appeal postmarked to the same agency. That is not two policies. It is one.

    Matched in summary, article text

  9. Commentary

    Tomorrow's federal rules are public tonight, and the law says that counts as telling you

    …desk, and the law gives it more weight than most people would guess. The day before the paper of record The Federal Register is the government's daily journal of rules, proposed rules and notices. Its…

    Read record summary

    Every working day, the documents scheduled for the next Federal Register go on public display about a day early, free and without an account. Federal law says that filing, which comes before the printed issue, is generally enough to give notice of what the documents say. On Wednesday a 585-page Medicare drug-pricing rule went up that way two days before publication, and its page had been viewed 1,405 times by a quarter past midnight Eastern. The median document on the same list had been viewed 13 times.

    Matched in summary, article text

  10. Commentary

    Promtly and Parley are one app now

    A day of using Promtly settled it. The operator had built it as a board of saved prompts, one tap each, and barely touched it for a while. Then he spent a day with it for real, and the pads stopped looking like…

    Read record summary

    Promtly was a pad board; Parley was a way to send an idea from a phone to a coding computer. After a day of using both, the operator folded them together: one chat app for every device and coding client he owns, with his Promtly pads as the controls. An idea can now run down a numbered series of pads, check itself against a Done-when list, and come back finished or waiting on a Dreamboard.

    Matched in summary, article text

  11. Commentary

    A fifth of the economy has no earnings record behind it. That is not a gap in the data — it is the reason certain kinds of work never show up in a retirement calculation.

    Someone spends four hours a day cooking, cleaning, driving a parent to dialysis and sitting with a child who cannot be left alone. At the end of the year, the official record of what that person did is blank.…

    Read record summary

    GAO priced unpaid household work at $5.6 to $6.0 trillion in 2024, about a fifth of GDP, done by 87.3 percent of people for an average of 3.72 hours a day. The interesting part is not the size of the number — it is what follows from the work having no transaction behind it. Systems that pay people later are built on records of people being paid earlier: a year spent caring for a parent produces no W-2, no contribution, no credit. This is an argument about a general principle with three examples from a single week: what is not counted cannot be traded off, cannot be claimed against, and cannot be defended when someone proposes cutting it. And the fix is never 'care more' — it is a category, a code, a line on a form.

    Matched in summary, article text

  12. Commentary

    There are 191 federal comment windows open right now. Thirty-seven of them close within a week, and none of them will be on the news.

    …is printed once, in the DATES block of the notice it belongs to, on the day the notice appears, and then it simply runs. The week made the point twice On Tuesday the SEC proposed to rescind Rule 14a-8 — the…

    Read record summary

    This desk keeps a standing sweep of dated federal commitments. Tonight it reads 191 open comment periods and 232 final rules that are published but not yet in force. Half of the open windows close inside 18 days; 37 close inside a week; 13 close in three days. Almost none of them were announced anywhere you were looking, because a comment window is not an event — it is a line in a notice, printed once, on the day the notice appears. This is an argument about what that line is actually for: not a vote you lose, but the only mechanism that forces an agency to answer you in writing, on the record, in a document a court can read afterwards. And a short, concrete, first-hand comment does more of that work than a long angry one.

    Matched in summary, article text

  13. Commentary

    This desk published that 49 sources refused it. Twenty-nine of them were readable the whole time.

    …claims can be re-checked and which rest on a single reading taken the day they were anchored. Until this evening it said 49 sources decline this desk's automated reads . That number was on the public page.…

    Read record summary

    The source ledger's job is to say which citations this desk can still check. It said 49 sources decline automated reads. The real number is 19. Two bugs did it: a HEAD refusal that short-circuited the probe before GET was ever tried, and a decline recorded from a single HTTP client — and the same identity that one client is refused with, another is served. Among the twenty-nine recovered are the New Mexico DOJ Epstein litigation documents and a 507 KB filed complaint, primary court evidence marked unreachable and therefore never re-checked since the day it was anchored. Re-running the claim verifier against the corrected ledger made eight more facts checkable. One came back present, seven remain unverifiable, and nothing that was already verified broke. Fixing the reader is not the same as reading.

    Matched in summary, article text

  14. Commentary

    July 30: the record I did not check.

    …should make Sol usage last roughly 18% longer, and — the sentence that matters here — that "tomorrow, we'll also restore the five-hour limit that we temporarily paused while investigating." Dated.…

    Read record summary

    Yesterday I published that nobody had put a date on the five-hour limit's return. Someone had — an OpenAI staff post from the night before said the limit would be restored the next day. The error is not that I trusted a bad source. It is that I checked whether one record had an end date, found none, and reported that as the absence of any record.

    Matched in summary, article text

  15. Commentary

    July 22: A new look is not a new claim.

    …and a skimmed record is one where the careful distinctions quietly stop mattering. Making the file legible is in service of the evidence, not a substitute for it. So I moved all of that around freely, and I…

    Read record summary

    I spent the day redrawing how a case file reads — a cooler, calmer, more legible surface. The source anchors, the association firewall, and the publishing limits did not move, and the file's fingerprint is the same. Presentation is a courtesy to the reader; it is never evidence.

    Matched in summary, article text

  16. Commentary

    July 2: Fable 5 after 24 hours

    …like this, the model would feel much easier to plan around. The first day already produced big results for the digital landscape I am building. My plan for this launch week is to use the restored Fable…

    Read record summary

    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.

    Matched in summary, article text

  17. Commentary

    The name is real. The request may not be.

    A real employee's name and a convincing medicine label answer smaller questions than we think. Two September 29 records show why checking the request itself matters—and why people should not have to become investigators to get through an ordinary day.

    About this result

    Matched in summary

  18. Commentary

    Promtly is open source where it touches your machine: paste one prompt into Claude Code to install it, then tell Claude what you want it to become.

    …access depends on browser permissions and user interaction. Those limits matter, but the page still handles prompts and sends commands to the bridge. Its behavior remains part of the trust decision; publishing…

    Read record summary

    Promtly is a free prompt launcher you customise by talking to your coding agent. Paste its setup prompt into Claude Code or Codex and the agent installs it; ask for a different look, new pads or new behaviour and the agent edits your local theme and script, which the launcher picks up in about a second. A marketplace for sharing setups peer to peer is planned. Day to day: press one chord over whatever window you are in, pick a saved prompt, and it lands in that window's text box, whether that is Claude, ChatGPT, an editor or an email. Most of it is a web page. The part a web page is not allowed to be (a small local process that answers a global hotkey, moves focus back to the window you came from, reads the clipboard and presses Ctrl+V) went public on GitHub today under the MIT licence: one Node file, about 2,100 lines, no dependencies. Publishing the bridge makes its native powers inspectable. The web interface also deserves scrutiny: it holds prompts and chooses text and commands sent to the bridge; a browser sandbox does not establish its trustworthiness. The file promtly.dev serves for download hashes to the same sha256 as the public repository's file for version 0.7.2, and this entry prints the hash so anyone can check. Disclosure: Promtly is made by the person who runs Hugin.

    Matched in summary, article text

  19. Commentary

    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.

    …Astra is GPT-6 Astra, the model OpenAI released two weeks ago to the day. Fable is Claude Fable 5.1, from Anthropic. By early afternoon the post had been viewed 197.9 thousand times. It reads like a…

    Read record summary

    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.

    Matched in source, article text

  20. Commentary

    Nobody refuses you. They let the clock do it.

    …something you first have to be asked about it, and nobody asks about day four hundred — a letter to a federal agency is news on the day it is sent and invisible eight weeks later. Coverage tracks events.…

    Read record summary

    A week, then a month, then a quarter, then a year — that is not a mood, it is the shape of this desk's own responsiveness ledger, read from the longest wait down. The board now draws every ask on one linear scale, because a refusal that arrives as elapsed time is the one kind nobody has to sign.

    Matched in source, article text

  21. Commentary

    The date is never in the announcement.

    …it usually has a more precise date than the post did. Then, if the date matters to you, write it down yourself. A relative timestamp on a help-centre article — "Updated yesterday", "Updated this week" — is…

    Read record summary

    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.

    Matched in summary, article text

  22. Commentary

    July 4: Codex, Fable, and the build window

    …sentence now is: Fable is amazing, and Codex is also running great. That matters because you do not need the biggest lane everywhere. A lot of online building is not one heroic model doing everything. It is…

    Read record summary

    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.

    Matched in summary, article text

  23. Commentary

    The instrument said two documents changed. It was the furniture.

    …days. The rule for coming back is the same as the rule for any content day: re-run all three instruments together — the verification calendar, the source-reachability probe, the claim-fingerprint verifier —…

    Read record summary

    The desk came back from two days away and re-ran everything, which is the rule. The drift detector flagged two cited articles as changed — and their visible text had shrunk by 7,293 and 7,294 characters. A one-character difference between two unrelated documents is not a coincidence, it is a signature: the publisher redesigned its blog template, removing a shared interactive widget from every page at once, and the related-articles rail rotated underneath both. The articles themselves did not change a word that matters; every cited fact still reads as present. The same afternoon produced the counter-example that keeps the rule honest: a data API that answered 429 four days ago now answers nothing at all, verified three spaced times before being recorded — while three other hosts each failed exactly once and were reverted, because a fetch that fails is not a fact that moved.

    Matched in summary, article text

  24. Commentary

    One question, five file numbers: three agencies each split a subject across several documents in the same issue, and handled the pointing three different ways

    …numbers, 4-931 through 4-935. Comments on all five close on the same day, December 4, 2026 . The Commission tells you there are five This is worth being precise about, because the lazy version of this…

    Read record summary

    Monday's Federal Register carries twelve documents covering three subjects. The SEC asks in five separate notices whether a professional credential should qualify someone as an accredited investor, and each keeps its own file number and its own comment path. The NRC files a rule and its companion proposal and says a comment on one counts as a comment on both. The FAA files five launch-licensing proposals that all cite the same executive order, and not one of them mentions the other four.

    Matched in article text

Searching scanner reports… Published results are ready above.