A running bit, traced to its source

Joey coin Eleven years of a joke about celebrity tokens — and the awkward discovery that it was never his joke

The bit is simple enough: anyone will be able to mint a coin, so there will be a Joey coin and a Freddie coin, kids will trade them in the playground, and sooner or later somebody will issue a token for a pop star. It runs across four of his shows from 2014 to 2025. This report was commissioned to establish the earliest instance of it in the archive. The earliest instance is not his. It is Andreas Antonopoulos’s, spoken on Thomas Hunt’s own programme on 30 January 2014, and Hunt has said so on air repeatedly — most recently in 2025. What is his, and what turns out to be ten months earlier than anything else here, is the Justin Bieber hook.

Archive material from 7,072 distinct broadcasts and 50,822,797 words across twenty-five searchable shows (WCN · Mad Bitcoins · The Bitcoin Group · The Bitcoin Talk Show · Today in Bitcoin · Brother John F · Ugly Old Goat · Let’s Talk Bitcoin · LTB Network · Epicenter · SF Bitcoin Meetup and fourteen more), 2011–2026, plus 91,295 tweets and 119,245 likes from the host’s own archive · 30 distinct broadcasts contain the phrase, 2014-01-30 to 2025-07-19 · Outside reporting used only in §09, kept separate · Compiled 2026-10-25, the day after 3,034 transcripts were upgraded to distil-large-v3 — see §07, which is why five of those thirty broadcasts are legible for the first time

01What this report was asked to establish, and why the answer changed

The commission was specific. The Curio book research had surfaced what looked like the earliest written instance of the Joey coin idea in the archive — Hunt’s own, in print, 30 March 2014, moving from corporate coins to celebrity coins to a Justin Bieber token — and this report was to establish that instance, hedge it properly as a statement about the corpus rather than about priority, trace the recurrences, and judge how the prediction fared.

The hedge was the right instinct and it is kept below, in §11. But it turned out to guard the wrong flank. The brief anticipated that someone might later turn up a January 2014 forum post and undercut a priority claim. What actually undercuts it is inside the archive, on Hunt’s own channel, and it is not a near miss: the phrase arrives eight weeks before the article, in a guest’s mouth, during a broadcast Hunt was hosting.

What survives, and is worth the report, is everything the brief said would stand on its own: that he articulated a direction and the industry went there. It just happens that he articulated somebody else’s direction, added a hook of his own that predates all of it, and then kept the argument alive for eleven years after the man who coined it had moved on. §10 is where that gets judged.

02The earliest instance in the archive — 30 January 2014

The query is corpus.py first '"joey coin"', phrase-quoted per CONVENTIONS §4d, because the unquoted form ANDs joey and coin across whole transcripts and is worthless for a date. It returns 2014-01-30, Mad Bitcoins, and 30 broadcasts in total.

The episode is MadBitcoins Live: New York Bitcoin Hearings — Day 2 (part 2), the second day of Hunt’s live coverage of the New York Department of Financial Services hearings. It is a 292,000-character transcript of a Google Hangout, and the Joey coin passage sits about two-thirds of the way in, inside a long uninterrupted monologue.

Two things about that passage deserve saying before anything is built on it.

It is not about celebrities. It is about children, and about the collapse of the cost of issuance. The example is a primary school, the prediction is explicitly that these coins will not have monetary value, and the argument it serves is a point about where monetary value comes from. The celebrity rung gets added later, by the same man, in March (§03). Reading the 30 January passage as an early celebrity-token prediction is reading the destination into the first step.

It is a prediction he is already presenting as settled. “As I’ve predicted in my more recent understanding of this field” is a back-reference. So even inside this archive, 30 January 2014 is where the phrase first becomes audible, not where the thought starts. §11 states what that does and does not license.

03Antonopoulos’s spring of 2014 — the same set piece, four shows, nine weeks

The 30 January passage is not a one-off remark. Between late January and early April 2014 the same argument appears on four different programmes, two of them Hunt’s and two of them other people’s, in progressively more developed form. Laid out in order it is obviously a stump speech being refined on the road.

1. 2014-01-30 — Mad Bitcoins (his)The bare version. “Joey coin and Jimmy coin and Maria coin in your primary school. And these will not have monetary value.” Children, no celebrities, no mechanism named.
2. c. February 2014 — The Bitcoin Group #16 (his)The full taxonomy, and the first appearance of the word celebrity. Also the clearest attribution in the whole report, because he is introduced by name and then identifies himself. See the quote below.
3. 2014-03-11 — LTB Network, Sovereign BTC #6 (not his)A full recorded lecture on scarcity and abundance. The mechanism is now costed — “a tenth of a bitcoin… there’s going to be a site that does this for free by the end of the year” — and the celebrity rung is explicit: “some of them will be Joey Coin and Maria Coin and Kanye West Coin.”
4. 2014-04-01 / 2014-04-02 — LTB Network and Let’s Talk Bitcoin #96 (not his)The polished line, delivered twice in two days: “When Joey can launch Joey Coin in his school with two clicks of a web interface in order to compete against Maria…”

The Bitcoin Group #16 is the pivotal one, and it is worth quoting at length because it contains the entire later history of the bit in a single turn — the playground, the cost of minting, and the word that the 30 March article would build on.

Note the shape of what he got right and wrong in that one turn, because §10 depends on it. The mechanism is almost exactly correct — ten minutes and a web form, which by 2024 is ten dollars on a Solana launchpad. The list is correct: brand, meme, fad, personal, promotion, reputation, celebrity. Every one of those exists now. The valuation is badly wrong, and in a specific way: he expects them to be worthless, or at most to carry “collectible value or simply reputation value.” Nobody in this archive predicts a personal token reaching a nominal fifteen billion dollars in two days.

And Hunt has been saying all of this himself, on air, for years

This is the part that makes the correction in §01 unembarrassing. The attribution is not a finding this report prised out of the transcripts against the subject’s account; it is his account, stated repeatedly and unprompted.

His recollection that it was “the first 10 episodes or so” of The Bitcoin Group is off by a few: the developed version is #16, and the earliest instance anywhere is the Mad Bitcoins hearings coverage rather than The Bitcoin Group at all. For a decade-old memory of which week a guest said something, that is close enough to be worth noting as accurate rather than as an error.

04What is his — Bieber, 29 May 2013

If the phrase is not his, the hook is. And it is the earliest thing in this report by a clear eight months.

Mad Bitcoins #031 is a short solo news episode — 756 words, one voice, signed off “this has been Mad Bitcoins” — and it ends on a speculative segment headed “what can cause the Bitcoin breakthrough?” Having worked through retailers and gas stations and rejected them, it arrives at artists.

But this is not the bit, and the distinction matters. The 2013 prediction is that a famous musician will accept Bitcoin and drag the public in behind them. There is no token, no issuance, no minting. It is a marketing-adoption theory, and as a prediction about Bitcoin’s route to the mainstream it was simply wrong — the mainstream arrived via exchange-traded funds and corporate treasuries, not via a pop star’s album.

What makes it matter is that it is the hook he already had in hand when Antonopoulos handed him the thesis eight months later. The fusion — celebrity from Antonopoulos, Bieber from himself — is what the 30 March 2014 article does, and it is a real contribution even though neither half of it is novel on its own.

The earliest written instance is a tweet, and it is five weeks before the article

So the sequence on his side is: the Bieber hook in May 2013; Antonopoulos’s thesis delivered on his own show in January and February 2014; “Biebercoin” in writing on 23 February 2014; and the article on 30 March. The article is the fourth step, not the first.

05The hedge, evidenced from inside the corpus

The brief asked for the claim to be framed as a statement about the archive rather than about priority, on the grounds that vanity and celebrity altcoins were already a visible genre by early 2014. That is correct, and it does not have to be asserted from general knowledge: the archive reports the genre as news while it is happening.

By the time Antonopoulos says “celebrity coins” on The Bitcoin Group #16, Kanye West’s lawyers had already killed Coinye — a cease-and-desist in January 2014 followed by a federal suit that same month, after which the creators declared the coin dead (press reports of the filing date differ between 8 and 15 January and this report does not resolve which is right). Maxcoin, promoted by Max Keiser, launched in early February 2014 and is discussed on #16 itself as a live story. These are not forecasts. They are the week’s news.

Three months later the term is being used on air as an established category, about a coin named for a film actor:

And by January 2015 the shows are filing the whole thing under last year’s fads, in a retrospective that also supplies the best single piece of evidence that the genre predates everything in §02 and §03:

The participants themselves could not remember which celebrity coin came first, in January 2015, about events twelve months earlier. That is the best possible argument for the framing in §11: nobody was first in any sense this archive can settle, and the archive’s own participants knew it at the time.

06The chronology — thirty broadcasts, and what the bit turned into

Order of adoption, from corpus.py spread '"joey coin"'. Six shows, de-duplicated to distinct broadcasts. Four are Hunt’s; two are not, and both of those are Antonopoulos talking.

By year, across the whole indexed corpus: 2014: 7 · 2018: 2 · 2019: 1 · 2020: 4 · 2021: 8 · 2022: 1 · 2023: 2 · 2024: 2 · 2025: 3. Nothing at all between October 2014 and July 2018 — a gap of nearly four years in which the phrase disappears from an archive that was broadcasting throughout. It comes back in 2018 and peaks in 2021, which is the NFT year.

The two peaks are the tell. The bit is not a steady catchphrase; it reactivates when the market starts minting things again. 2014 is the altcoin wave, 2021 is NFTs and social tokens, 2024–2025 is Solana celebrity memecoins. In between it goes quiet.

What the bit came to mean — four distinct uses, in order

Tracking the phrase across eleven years shows it changing job three times. The words stay the same; the argument underneath them is replaced.

1. 2014 — a thesis about abundanceAntonopoulos’s original: issuance is free, so coins proliferate without limit, and monetary value becomes an emergent property of adoption rather than a property of the issuer. “These will not have monetary value.” The claim is about where value comes from.
2. 2014–2019 — a slot in a listIt flattens into a generic altcoin name. Bitcoin Talk Show #22, 20 August 2014: “…like coin, feather coin, very coin, viac coin, meac coin, Joey coin, Fredi coin, Ralph coin. All of these coins are…” By July 2018 it is a routing problem — “even if it’s Joey coin, Freddie coin, Target coin, there’s going to be all these hops.” And on The Bitcoin Group #196, 22 March 2019, it is simply a punchline: “Why use Lightning for tipping when you could use tipping coin or Joey coin? Joey coin, a very popular coin developed by…”
3. 2021–2023 — an access token, and a Curio Cards lineageIn the NFT years it becomes a worked example of a personal token with a use. The Adventures in NFTs interviews return to it three times across #3, #4 and #5 in a fortnight; the Curio Cards sixth-anniversary retrospective of 10 May 2023 puts it as “if Joey wants to have a party, and say you have to have Joey coin to go to the party, Joey can…” This is the phase where the bit is doing intellectual work for him again, and it is also where he says it influenced Curio Cards.
4. 2022–2025 — a byword for something worthlessThe last turn is the bleakest. The Bitcoin Group #308, 20 May 2022, during the Terra/Luna collapse: “people look at these things and they’re like, I was up on, you know, Joey coin up 4,000 percent — but if you would try to sell Joey coin at the top there wouldn’t have been enough buyers.” Joey coin is now a liquidity lesson. The optimistic thesis has been inverted into a warning, using the same three syllables.

Stage 4 is the honest end of the arc, and it is worth being clear that this is a decay rather than a development. In 2014 the phrase carried an argument that minting would become free and that this was fine. By 2022 it names the thing you cannot sell. The 2025 instances (§10) partially recover the argument, but only in order to say it went wrong.

07What the transcript upgrade actually added — measured, not assumed

This report was compiled the morning after 3,034 transcripts were re-cut from whisper-base to distil-large-v3. The brief said passages that were not searchable the day before would be searchable now. That is true here, and it is checkable rather than merely plausible, because every upgraded row keeps a transcript_superseded record pointing at the file it replaced.

So the check is exact: count the phrase in the current text, count it in the previous file, and compare. Of the 41 rows carrying the phrase, 13 are upgraded rows. Five of those contained no match at all before the upgrade.

The consequence is substantive rather than cosmetic. Every quotation in §10 — the section that judges whether the prediction came true — comes from a row that was illegible for this phrase until yesterday. A report written on 23 October would have had the 2014 material and the 2021 material and would have had to guess at the reckoning, because the four broadcasts in which he actually delivers it were invisible. That is the whole argument for re-running old queries after an upgrade, and §08 is what it cost the previous count.

08Counting it properly — 41 rows, 30 broadcasts, and a rebroadcast

Three different numbers are defensible depending on what is being counted, and only one of them belongs in a sentence about how often he said this.

The rebroadcast, because it would otherwise read as a recurrence

spread reports Mad Bitcoins carrying the bit three times, 30 January to 21 October 2014, and that looks like a host returning to a favourite line over nine months. It is not. The transcripts of the 20 October and 21 October 2014 MadBitcoinsTV livestreams contain Antonopoulos’s January monologue verbatim — the four-football-field printing press, the unwashed masses, the five-year-old with a GUI, all of it, in the same order, differing only in Whisper’s errors. The October streams replay the January hearings coverage.

This is useful in one way and misleading in another. Useful: two independent transcriptions of the same audio agree on the wording, which is a real corroboration of the §02 quote. Misleading: it inflates the show’s count threefold and extends its apparent span from one day to 264. Mad Bitcoins has one original instance of this bit, on 30 January 2014.

And why this disagrees with the repertoire report

The recurring-repertoire survey local archive only — _local/repertoire/ is not deployed, so there is no public URL to link catalogues the Joey coin bit at 20 broadcasts, 2014-01-30 to 2025-01-25, across four shows. It does not analyse the bit; it records the count, notes that a dedicated write-up was to follow, offers its own figures only as an independent cross-check against that write-up, and states that where the two disagree the dedicated write-up — the one that read the passages — governs. They disagree. This is that write-up, and the difference resolves cleanly into two causes.

The residue after both causes is small and is not fully reconciled here; the repertoire figure was derived by a different sweep whose exact query is not recorded on its page. The 30 in this report is reproducible — corpus.py spread '"joey coin"', phrase-quoted, on the index of 25 October 2026 — and the repertoire page should be updated to point here rather than the two being averaged.

09What actually happened — outside the archive

Everything from here to the end of this section is outside reporting, not archive material, and is listed separately in the sources. Nothing in it establishes anything about what was said on air.

Before leaving the archive, though, one find belongs at the head of this section, because it settles a question the brief could not have known to ask: was anybody actually building the mechanism in 2014? Yes, and they were using Bieber as the worked example, on the LTB Network, seven months after Hunt’s tweet.

A month later, on Let’s Talk Bitcoin #164, the same model is described in full and the reasoning is exactly the fan-token logic of six years later: “you can create your own artist coin. So Snoop coin or Bieber coin, whatever coin you want. … your fans own a piece of your coin, [so] they now have an incentive to promote you.” So the idea was not merely in the air in 2014 — it was in development, with a pitch deck, and the pitch named Bieber. Neither man in this report was predicting something nobody was attempting.

The genre, dated

What follows is the record from contemporaneous press, in order. Figures from crypto-market reporting are given as reported and should be read as indicative: market capitalisations for tokens of this kind are nominal, move by orders of magnitude within days, and the sources rarely state the hour of measurement.

10Was he right? — right about the shape, wrong about the direction

The brief asked for an honest answer, including where the answer is no. It turns out the best available answer is his own, delivered on 25 January 2025 eight days after the TRUMP launch, and it is more critical of the prediction than this report would have been on its own.

That is the verdict, and it divides cleanly into three findings.

Right about the mechanism, in detail

Antonopoulos’s February 2014 claim was that a five-year-old could make a coin “in about 10 minutes flat” and that a service costing a tenth of a bitcoin would be free within the year. By June 2024 the figure quoted on The Bitcoin Group is “for about ten dollars I think you could launch Joey coin” — and DADDY was in fact launched through a one-click launchpad. The cost-of-issuance prediction is simply correct, and it was correct about something non-obvious: that the cost rather than the technology was the binding constraint. His taxonomy — brand, meme, fad, personal, promotion, reputation, celebrity — now reads like a list of market segments.

Wrong about who would use it, and he says so

Both men described a bottom-up phenomenon: playgrounds, five-year-olds, rubber bands and Pokemon cards, “young hominids” inventing currencies among themselves. Hunt’s 2025 gloss extends it to a sort of personal finance — mint coins to get through college or a hard year. None of that happened. The tokens that mattered were issued by or for a head of state, a rapper, and an internet celebrity, with insider wallets taking a fifth of the supply before the promotion started. The distributional prediction inverted.

It is worth being exact about why that is a real miss rather than a quibble. The 2014 argument was not just a forecast that many coins would exist; it was a claim about what cheap issuance means — that value would become emergent, communal and bottom-up, decided by users rather than issuers. The mechanism arrived and delivered the opposite: issuance stayed cheap, and the value accrued to whoever already had the audience. Cheap tools did not distribute power.

Wrong about Bieber, specifically and completely

The named prediction — from 2013, from the February 2014 tweet, and from the 30 March 2014 article — was a Justin Bieber token. Justin Bieber has never issued one. His involvement ran the other way: he bought a Bored Ape in January 2022 for about $1.3m and it was reported at around $12,000 three years later. The only “Justin Bieber coin” that ever traded was minted for him without his consent by BitClout in 2021 — a mechanism nobody in this archive anticipated, and one that makes the celebrity a victim of the token rather than its issuer.

Against that, Antonopoulos’s own 2014 example fared better. He said “Kanye West Coin” on 11 March 2014 — two months after Kanye West had sued Coinye out of existence. He was describing something that had already happened and been litigated, which is easier than predicting, but his choice of name was the one that stuck: celebrity coins in the end looked far more like Coinye than like Joey coin.

The counter-argument, made on air at the time

The report should not leave Hunt’s verdict unchallenged, because his own co-panellist challenged it in the next breath and the objection is a good one.

Segalo’s point is that Dogecoin was genuinely bottom-up and that the celebrity tokens are a late capture of a grassroots phenomenon, on the ICO pattern — which, if right, means the 2014 prediction was correct and merely got overrun. The disagreement is not resolved on air and this report does not resolve it either. It does observe that the two readings make different forecasts: on Hunt’s, the personal coin never arrives; on Segalo’s, it is still coming and will be small. As of this compilation, eighteen months later, there is no evidence in the archive or outside it of a general market in ordinary people’s personal tokens.

What stands without any of it

Strip out the priority claim and what remains is the thing the brief said would stand on its own, and it does. In early 2014 two people on a live stream about a New York regulatory hearing worked out that the cost of issuing money was going to zero, that the result would be brand and meme and celebrity coins, and that the interesting question was where value would come from once issuance was free. Twelve years later a sitting US president launched a token and it was briefly worth more than most national currencies — which is close to the sentence Antonopoulos was already using in January 2015, when he told an SF Bitcoin Meetup audience that “Justin Bieber coin, at this point, would be worth more than about 30 national currencies if it launched tomorrow.”

He was not predicting first. He was predicting accurately, in public, on the record, in an archive that can now be searched to prove it. That needs no claim about who said it first, which is just as well, because the archive says he did not.

11Coverage — what these dates are claims about

Per CORPUS.md §5b and the brief’s own instruction, the coverage statement goes in the report rather than in a session log. Every date in this report is a claim about what is currently indexed, not a claim about history.

Two further limits that are specific to this subject rather than general.

The phrase is a bare two tokens, and one of them is a common word. CONVENTIONS §4d and CORPUS.md §5 both require snippet-checking anything this short. Every one of the 41 rows was snippet-checked and every hit is the bit; there are no false positives of the "a i" kind. But the inverse is not covered: a passage where he makes the same argument without the words “Joey coin” does not appear in any count here. The variant sweep (joeycoin, joey coins, jimmy coin, maria coin, celebrity coin, celebrity coins, vanity coin, bieber, justin bieber) is the mitigation and it is not a substitute.

The 30 January 2014 row is a whisper-base transcript with a repetition-loop flag. It has not been upgraded. Its wording is corroborated by the independent distil-large-v3 transcript of the October rebroadcast (§08), which is why the quote is used; but when the distil pass reaches it, this report’s central quote should be re-verified, and the two dozen whisper-base rows in §06’s chronology re-counted. §07 shows that upgrades move counts in both directions.

Re-run before relying on any figure here. The reproducible commands are corpus.py first '"joey coin"' and corpus.py spread '"joey coin"', run from _scripts/corpus. Both print their own coverage block, which is the one to read; the block above will go stale the moment any of the ten held shows is ingested.

12Sources