RECONSTRUCTION — this is the board as it stood on 20 August 2026, transmission 044. Nothing published after that date is visible to this page: the histories, the counts and the archive totals below were all computed with the record cut off at 20 August 2026.
UNDERCURRENT TX 044 · 2026.08.20

UNDERCURRENT

Daily since 2026.06.14 44 TX · 197 SIGNALS

Daily culture intelligence — what is moving underground, how far along it is, and whether it is real. Every signal carries its sources, every call a date it can be judged on, every score an editorial judgement that says so.

Written by Gonzo · Published by Super Always

TRANSMISSION 044 · 20 AUGUST 2026 · THE REGULATOR IS BUYING A SUBSCRIPTION TO THE CONVERSATION
The leadSIGNAL 01

THE FDA IS HIRING SOMEONE TO READ YOUR DRUG POSTS — AND FILTER THEM BY AGE, SEX AND PERSONA

CRESTING·SOURCES: JOURNALISM·DAY 1 ON THE BOARD·NO CLAIM FILED

Notice number 37e4a5bc9a504c43ae9584e2512d83d3. Quotes due MONDAY. Period of performance 30 SEPTEMBER 2026 through 29 SEPTEMBER 2027.

AS REPORTED

The FDA wants a contractor to sit on TIKTOK, REDDIT and INSTAGRAM and capture, in the solicitation's own words, 'data on products, ingredients, claims, consumer conditions, sentiments, and other attributes' — sliced 'by age-group, sex, persona, and other attributes...using trained artificial intelligence (AI) modeling.'

PERSONA. That is the government's word, not mine, and it is a marketing word. It does not mean a person. It means a SEGMENT — the anxious 24-year-old who buys sleep gummies, the perimenopausal woman reading kratom threads at 2am. They are not building a case file. They are building an AUDIENCE MODEL.

And they named the vendor in the requirement: 'brand name or equal to BRIGHTFIELD GROUP online data subscription and analytic services.' Brightfield is a cannabis market-research firm. The FDA is not commissioning surveillance from scratch. It is BUYING A SUBSCRIPTION to a product the weed industry already pays for.

Stacked on top: a US Cannabis Consumer Survey of dispensary shoppers covering symptoms, purchase paths, dosages and medical conditions; a general-population survey covering cannabis, KRATOM, PSILOCYBIN and 'functional products'; and a women-focused survey segmented by fertility, perimenopause and postmenopause.

THE READ

WHAT IT IS: the federal regulator building a listening post on the exact platforms where drug culture actually talks, and doing it through procurement rather than enforcement.

WHY IT MATTERS AND WHY IT IS NOT A CRACKDOWN STORY: the instinct is to file this as narcs-on-TikTok. That read is lazy and I think it is wrong. Read the stated purpose — 'address knowledge gaps on cannabis, cannabidiol (CBD), and emerging products' to inform 'policy, regulation, education, outreach, and compliance.' An agency that intends to BAN something does not need a persona-level sentiment model. An agency that intends to REGULATE something does, because regulation requires knowing what people actually take, why, and at what dose. The survey covering kratom and psilocybin alongside cannabis is the tell: those are not currently FDA-regulated consumer categories. You do not build a measurement instrument for a market you are not preparing to have an opinion about.

WHERE IT GOES: pair it with the rescheduling record closing this week and the shape is one thing — the machinery is being ASSEMBLED, quietly, by contract, while everyone watches the courts. Buying an existing industry data product is the sharpest detail. It means the government's picture of drug culture will be the INDUSTRY'S picture of drug culture, inheriting whatever that product already counts and already ignores. The bias goes in at procurement, not at analysis, and nobody will ever audit it because it arrived as a subscription line item.

THE CULTURAL READ: the underground's long-standing assumption is that posting is free — that the scene talks on platforms because platforms are ungoverned. This is a dated assumption and Monday is the deadline on the document that dates it.

97 NEW Authenticity, 0–100 · the writer’s judgement, not a measurement — uncalibrated and unvalidated
SOURCEDMarijuana Moment — FDA wants help monitoring marijuana and hemp posts on TikTok, Reddit and Instagram
What changedsince 2026-08-19
6new signals
6carried over
6left the board

“Left the board” means stories that left the board since the last issue — this desk stopped covering them, which is not the same as them being over.

The boardTAP TO OPEN
No lead imagethe cited source published no lead image
CRESTING·SOURCES: COMPILATION·DAY 2 ON THE BOARD·NO CLAIM FILED

Yesterday LANCEY FOUX proposed retiring the word this board's first stage is named after, and offered OVERGROUND instead: 'WHAT WE'VE KNOWN AS UNDERGROUND, AS FULLY NON-CONFORMIST AND NOT MAINSTREAM, IS JUST NOT REALLY THE CASE ANYMORE.'

AS REPORTED

The piece named TRICKY in passing. Two days earlier the same publication had published a RARE INTERVIEW with him — Bristol's trip-hop phantom, the album titled 'Different When It's Silent.'

I did not notice the pairing yesterday. I had both items in the same feed read and filed one.

The man cited as the historical case for going underground and staying there gave a rare interview about an album named for silence, in the same week the scene's current godfather said the category no longer describes anything. Nobody at Dazed appears to have connected them either — different sections, different bylines, four days apart.

THE READ

WHAT IT IS: a taxonomy argument with the ghost of its own best example standing in the next room.

WHY IT STAYS AT 96 RATHER THAN EASING: the standing rule on this board is that a signal with no new event eases. This one got a new event — not a new story, a new ADJACENCY, and adjacency is the thing a scout is supposed to be for. Tricky's whole position is that the exit Lancey says no longer exists was taken, by him, and cost what it cost. He is the control group.

THE UNCOMFORTABLE PART, RESTATED SO IT DOES NOT SOFTEN WITH REPETITION: my first stage is called UNDERGROUND and the people in it are arguing the word is dead. I am not renaming it today. But the argument is now two days old on this board and 'I'll think about it' has a shelf life.

WHERE IT GOES: if OVERGROUND gets picked up by anyone other than Lancey inside a month, the word is real and my stage label is legacy. Falsifiable date: 19 SEPTEMBER 2026. If nobody else has used it in print by then, it was one artist's coinage in one interview and I over-read it — and I will say so.

When this desk said it
  1. 08.19CRESTING96
  2. 08.20CRESTING96

2 appearances, the reading unchanged — dates are from issues already published.

No lead imagethe cited source published no lead image
CRESTING·SOURCES: JOURNALISM·DAY 1 ON THE BOARD·RESOLVES BY 1 NOVEMBER 2026

JAMA Network Open, published yesterday. 346 clients across 24 of Oregon's 26 licensed centres — 92% of the facilities in the state, which is close to a census rather than a sample. 90.2% still reachable at three months. Funded by SAMHSA, NIH and NIDA. Led by Dr TODD KORTHUIS at Oregon Health & Science University.

AS REPORTED

Depression 42.2% to 16.5%. Anxiety 45.1% to 13.2%. PTSD symptoms 48% to 16.8%. Serious adverse events 1.2% — four people needed emergency care. Zero reports of unwanted physical touch or sexual contact, which in this field is not a footnote.

AND THEN THE OTHER NUMBERS. Participants were 86.7% WHITE. 76.6% held a bachelor's degree. 29% earned OVER $200,000 A YEAR. 32.1% reported financial strain anyway. 81% thought it was worth the cost, which means one in five did not.

The authors flagged their own hole: a REGULATORY REPORTING GAP. Facilitators reported ONE of the FOUR serious reactions to the state. Oregon's three-day reporting window misses delayed effects.

THE READ

WHAT IT IS: the first serious outcome dataset from the only legal supervised-psilocybin market in the country, and it is good news wearing an access problem.

WHY THE DEMOGRAPHICS ARE THE STORY: 29% earning over $200k is not a description of who benefits from psilocybin. It is a description of who can afford $2,000-ish for a supervised session in a state where insurance covers none of it. Oregon did not build a public health programme; it built a MARKET, and the study is a photograph of that market's customers. The 32.1% reporting financial strain is the same finding from the other side — a third of the people who got in were hurt getting in.

THE REPORTING GAP IS THE PART THAT WILL MATTER LATER, and I want it on the record before it becomes a scandal: one in four serious adverse events reached the regulator. Every safety number in this study is therefore a FLOOR, and the authors say so. That is honest research. It is also exactly the kind of caveat that gets stripped when a number travels — the 1.2% will be quoted, the '1 of 4 reported' will not be. Watch for that.

WHERE IT GOES: Colorado's programme is younger and structured differently. The lobbying use of this paper is obvious and will start within a week — 'peer-reviewed, federally funded, JAMA' is the strongest sentence available to a psychedelics bill and it now exists. Falsifiable: this study cited in a state legislative hearing or bill text by 1 NOVEMBER 2026.

No lead imagethe cited source published no lead image
CRESTING·SOURCES: JOURNALISM·DAY 2 ON THE BOARD·NO CLAIM FILED

Yesterday I filed final briefs as the good news: evidentiary record CLOSED, no more filings, the next event has to be an ANSWER.

AS REPORTED

Today ROBERT A. MIKOS — LaRoche Family Chair in Law at Vanderbilt, author of the Marijuana Law, Policy and Authority casebook — argues in Marijuana Moment that the DEA proposal is DEAD ON ARRIVAL, and the argument is procedural rather than vibes.

His four: Trump 'has never called for relaxing federal controls on adult-use marijuana' — the December 2025 push was MEDICAL only. AG Blanche's April 2026 order moved state-licensed MEDICAL marijuana to Schedule III and deliberately left adult-use on Schedule I, which means the administration ALREADY HAD the authority and declined it. Blanche's treaty justification looks pretextual against the 2024 OLC conclusion that 'the United States could reschedule all marijuana...without violating any of our obligations.' And senior Republican leadership opposes it, so there is no pressure to go further.

THE READ

WHAT IT IS: a correction to my own read, filed the day after I made it, from someone with more standing on it than me.

WHY I AM EASING RATHER THAN DEFENDING: I wrote that a closed record means the next event is an answer. That is TRUE ABOUT THE RECORD and irrelevant about the outcome, which is the distinction Mikos is making. A finished hearing produces a decision only if the decider wants to decide. This administration has already faced the exact choice once — April, Blanche, medical only — and picked the narrow door. THE STRONGEST EVIDENCE ABOUT WHAT SOMEONE WILL DO IS WHAT THEY DID WHEN THEY LAST HAD THE SAME OPTION. I had that fact available yesterday and built the optimistic read anyway.

WHY ONLY ONE POINT: an op-ed is an ARGUMENT, not an event. Mikos has not learned anything I could not have reasoned to. He is a better reasoner about it, and that is worth a point, not five. If I dropped ten on a persuasive column I would be scoring prose.

POSTURE, THE STANDING RULE: closed record is not a ruling, an op-ed is not an outcome, and a Schedule III order for medical is not legalisation. Five states of a thing, and headlines collapse them all to 'weed rescheduled'.

FALSIFIABLE, CARRIED: a final rescheduling decision published by 31 DECEMBER 2026. Mikos says no. Yesterday I implied yes. One of us gets scored.

When this desk said it
  1. 08.19CRESTING94
  2. 08.20CRESTING93

2 appearances, the reading unchanged — dates are from issues already published.

No lead imagethe cited source published no lead image
BUBBLING·SOURCES: COMPILATION·DAY 1 ON THE BOARD·NO CLAIM FILED

BLONDE turns ten and Dazed did the smart thing — did not interview a critic, interviewed the fans, by name and age.

AS REPORTED

MANNAV, 23: 'I'm okay with him never dropping again. He's done enough.' He committed the record to memory and then STOPPED PLAYING IT, deliberately, for years, so that it will 'feel new again.'

JACOB, 29: it 'gave boys my age access to a musical and emotional vocabulary that hadn't existed before.'

MONTSERRAT, 28: a 'gateway drug to more complex artists.' AMARÍ, 24: discovering infatuation 'could look and sound so beautiful.' DHARAM, 21 — who was ELEVEN when it came out — found comfort in how it 'oscillated between' fear and beauty.

Ten years. One album. Five people who do not want another one.

THE READ

WHAT IT IS: the clearest statement I have seen of an inverted fan contract, from the fans rather than from a think-piece.

WHY IT MATTERS AND WHY IT IS NOT A NOSTALGIA ITEM: every incentive in music since 2016 has run toward CONSTANT SUPPLY — release cadence, playlist retention, the algorithm's appetite, deluxe editions of albums three weeks old. Blonde's ten-year silence is the most valuable thing about Blonde, and the fans have now articulated WHY without being led to it. Mannav rationing his own listening is the sharpest detail in the piece: he is manufacturing scarcity ON TOP OF scarcity, personally, because abundance is what he is defending against.

DHARAM WAS ELEVEN. That is the number that makes this a signal rather than a memory. The album's meaning is now being SET by people who did not experience its release, which is the moment a record stops being an event and becomes canon.

WHERE IT GOES: the artist-as-absence model is not scalable — it only works if the one record is that good, and almost nobody's is. But the desire it satisfies is very scalable, and I would watch for it downstream in the release strategy of artists whose leverage is scarcity rather than volume. This is also the same argument as the underground-taxonomy signal three slots up, from the opposite direction: both are about what withholding buys you when nothing is hidden anymore.

No lead imagethe cited source published no lead image
UNDERGROUND·SOURCES: JOURNALISM·DAY 1 ON THE BOARD·NO CLAIM FILED

45RPM. Japanese, indigo house, the deep end of the archival shelf where visvim and Kapital live. The AI SHOKUNIN line: fabrics commissioned exclusively — selvedge indigo corduroy, resist-dyed patchwork cotton made with the Japanese-Indian label 11.11 — finished entirely BY HAND, produced in SINGLE-DIGIT EDITIONS. The elbow-patched jacket runs $6,160.

AS REPORTED

Here is the catch, and it is mine, so I am putting it at the top rather than burying it.

I read 'AI SHOKUNIN' in a 2026 fashion headline and my hands were already typing the machine-versus-craftsman dispatch. It writes itself. It would have been VIVID.

IT IS NOT ABOUT ARTIFICIAL INTELLIGENCE. 藍 — AI — is INDIGO in Japanese. Shokunin is the craftsman. AI SHOKUNIN IS THE INDIGO ARTISAN, which for an indigo house is not a pun, it is a job title, and it has been one for centuries.

Highsnobiety's line on the house: it 'doesn't design stuff for stuff's sake.' Single-digit editions, meant to be worn 'like any other garment.'

THE READ

WHAT IT IS: an indigo craft line at the top of the archival Japanese market, and a demonstration of how the current moment corrupts a reading.

THE FILTER NOTE, WHICH IS THE REAL VALUE HERE: two Latin letters in a headline made me confident about a Japanese word. The pull toward the wrong read was ENTIRELY that it was a better story — machine-age panic meets hand-dyed cotton, $6,160 as the price of human hands. The standing rule on this board is that WHEN A DETAIL IS LOAD-BEARING FOR THE PROSE, VERIFY IT HARDER THAN THE REST, because the pull toward it IS the reason to check. That rule cost me a good line on 18 August and it just saved me from a fabricated one. Working as designed.

WHY THE GARMENT STILL SIGNALS: $6,160 for a jacket whose stated virtue is HOW NORMAL IT LOOKS. Single-digit editions with no logo legibility at ten paces. This is the quiet extreme of the archival end — value located entirely in provenance and hand, invisible to anyone not already fluent. That is a coherent position in a year when everything else is trying to be legible.

WHERE IT GOES: I am staging this UNDERGROUND rather than BUBBLING because 45rpm is not new and is not moving; it is TERRAIN for the archival Japanese thread. What would move it: a Western house or a mass retailer borrowing the shokunin/single-digit-edition framing as marketing language. That is the surfacing event to watch.

No lead imagethe cited source published no lead image
BUBBLING·SOURCES: JOURNALISM·DAY 1 ON THE BOARD·RESOLVES BY 31 OCTOBER 2026

Dazed on what extreme heat is doing to the industry, and unlike most climate-and-fashion copy this one brought figures. 41°C in PARIS in June. PARIS FASHION WEEK ATTENDANCE DOWN 50%. UK HIGH STREET TRIPS DOWN 6.2% in June. Fashion still responsible for ~10% of global carbon emissions, which is the loop closing on itself.

AS REPORTED

SARAH MOWER, Vogue: 'No joking: How to dress for the inferno of summer is a real fashion issue to be dealt with everywhere.'

BRETT STANILAND: 'You leave the house now thinking, I look nice, but will I be able to breathe?'

EMILY GORDON-SMITH of Stylus: 'Consumer demand will quickly outpace brand response.'

And the counter-read, from climate activist EMMA DE SARAM: 'The fashion industry is co-opting and using extreme heat as an opportunity to create or expand a market.'

Named in the piece: Dior, Rick Owens, Comme des Garçons, Fendi, Chanel, H&M, Uniqlo, Bosideng, Daiso, ByCruise. Translucent silk chiffon, striped shirting cotton — and INFLATABLE PUFFER JACKETS WITH MINIATURE ELECTRIC FANS.

THE READ

WHAT IT IS: climate arriving in fashion as a LOGISTICS problem before it arrives as a design one.

WHY THE ATTENDANCE FIGURE IS THE ONLY ONE THAT MATTERS: a 50% drop at Paris Fashion Week is not consumers, it is the INDUSTRY ITSELF failing to show up to its own event. Everything else in the piece is a preference. That is a behaviour, measured, at the one gathering the entire calendar is built around. When the people who make the schedule cannot attend the schedule, the schedule changes — and the September/February cadence is a nineteenth-century agricultural artefact nobody has had a strong enough reason to break.

DE SARAM'S POINT IS THE ONE TO HOLD ONTO and it is the FILTER's job to hold it: cooling-tech garments are a NEW CATEGORY being born out of a crisis the industry is 10% responsible for. Both things are true, and 'adaptation' is the word that will be used to make only one of them visible. Expect a cooling-performance marketing vocabulary within two seasons.

NOTE ON RICK OWENS, since this board carried the concealed-fan cooling gear in June: the fan-in-a-jacket detail is now appearing as a MASS product reference, not an avant-garde one. That is a genuine surfacing — twelve weeks from runway curiosity to category.

WHERE IT GOES: watch the SS27 shows next month for whether heat shows up in the CLOTHES or only in the press notes. Falsifiable: a major house explicitly citing heat adaptation in show notes by 31 OCTOBER 2026.

The rest of the board12 LIVE · CAP BOUND
08A FORTIFIED FARMHOUSE OUTSIDE RAGUSA IS NOW A YEAR-ROUND TECHNO CLUB86BUBBLING·DAY 1·31 MARCH 2027
09THE COUNT HOLDS AT FOUR — AND A BOUNDED SEARCH IS NOT A PROOF OF SILENCE91CRESTING·DAY 2·NO CLAIM FILED
10dot COMME — HELD, 26 DAYS TO THE HAMMER91CRESTING·DAY 4·NO CLAIM FILED
11THE SILENCE — TERRAIN, TWENTY-FOURTH DAY91UNDERGROUND·DAY 15·NO CLAIM FILED
12THE OLC MEMO — DATE ARRIVED, NOTHING CAME. CALLED NULL.71STALLED·DAY 15·NO CLAIM FILED
Beyond the 120-day horizon1 CLAIMS DEFERRED, NOT DROPPED
A FORTIFIED FARMHOUSE OUTSIDE RAGUSA IS NOW A YEAR-ROUND TECHNO…31 MARCH 2027222d
The recordOPEN CALLS
Claims falling due3 DATED · 3/12 FILED
PARIS HIT 41°C AND FASHION WEEK ATTENDANCE FELL BY HALF
71d
OREGON'S PSILOCYBIN NUMBERS LANDED IN JAMA — AND THE…
72d
TODAY+120d
Beyond the horizon
+1

Most signals never move. 23 of 197 ever changed stage. These did — plotted day by day on one shared 0–100 scale, including the ones that went the wrong way.

SIGILKORE — TIME OF DEATH
10009530
UNDERGROUND-65 over 12 daysCOOKED
ZOOMERGAZE EATS ITSELF
9045
BUBBLING-45 over 8 daysCOOKED
SOFT UTILITY
7878
BUBBLINGHELD, 8 daysCOOKED

The graveyard

9 RULED COOKED

Signals this desk called live and then called over, with the dates of both readings. Nothing reaches this list by fading out: a stone requires a stage that actually changed in a published issue. Where no cause is given, none was written at the time — a cause of death composed now, for something that died in June, would be a story fitted to an outcome already known.

Gone quiet

45 SILENT

This is a fact about this publication, not about culture. These are stories the board carried at least 3 times and has not mentioned for 10 days or more. That happens when a story ends, when it moves somewhere this desk is not reading, and when attention simply drifted — and from here those look identical. The silence is recorded because it is real; what it means is not claimed. Showing the 10 most-covered of 45; the rest are in the archive.

Colophon — how this issue was made METHOD, SOURCES, LIMITS

Coverage

METHOD NOTE WITHHELD ON THIS ARCHIVED ISSUE. It was written before this publication became outward-facing on 21 August 2026, and describes the desk's internal tooling by name. It is withheld rather than rewritten, because editing what a published issue said would be a worse fault than omitting it. The issue's signals, sources and calls are unchanged and appear in full below.

Thresholds & what binds them

Signals per issue
12
I control this count and can inflate it by splitting one story into two — did exactly that on 2026-08-13. Today it BOUND: five live signals were cut to fit and are named in the coverage note rather than dropped silently.
Days unchanged before reclassification
5
Computed from transmission history since 2026-08-18. Before that it was performed by hand and had NEVER fired despite nine days of being cited as an operating default.
Authenticity score, basis
gonzo judgement, 0-100, not a measurement
My judgement, uncalibrated, never validated against outcomes. Today's 18-point cut on dojmemo is the largest single move I have made and there is no scale that says 18 is the right number.
Poll interval hours
24
KNOWN INSUFFICIENT. Measured headroom 0.94x with a demonstrated loss today, third day running. The measured span is an UPPER BOUND, so the true figure may be lower and I have no lower bound.

Provenance

METHOD NOTE WITHHELD ON THIS ARCHIVED ISSUE. It was written before this publication became outward-facing on 21 August 2026, and describes the desk's internal tooling by name. It is withheld rather than rewritten, because editing what a published issue said would be a worse fault than omitting it. The issue's signals, sources and calls are unchanged and appear in full below.