1n2.org · TBG ARCHIVE PROJECT
Home · TBG Mirrors · Saylor Report · Sanctions Report

The Bitcoin Group · the view-count numbers.

Twelve years. 482 episodes. One snapshot of YouTube view counts pulled from the canonical archive on disk. This is what the numbers say about the show's audience trajectory — peak year, current floor, and whether the audience that's still here is more engaged than the one that left.

Peak year median
2017 · 47 episodes
2025 median
40 episodes
Drop from peak
−95.3% in median
Total eps analyzed
919,008 combined views
Section I · The trend line

Median views per episode, year by year.

Each bar is the median view count for episodes aired that year. The mean is plotted on the secondary axis. 2026 is partial (9 episodes through March 7).

Median view count, by year aired

Aggregated across both Mad Bitcoins and World Crypto Network YouTube uploads per episode.
SOURCE: wcn-podcast/data.json snapshot, 2026-03-18 · n=482 canonical episodes

Every episode, plotted

Each dot is one episode. Y-axis is on a log scale to keep the 2017 peak and the 2026 floor on the same chart.
SOURCE: wcn-podcast/data.json · log scale to render 14–16,635 view range
Section II · The peak

2017 is the entire story.

2017 accounts for 423,125 of the show's 919,008 total recorded views — nearly half the archive's lifetime views came from a single year. The top 10 episodes of all time are all from a five-month window between July 2017 and January 2018.

EpDateViewsTitle

Every spike is a 2017 bull-run news beat. The panel didn't cause the spikes — the price did.

Section III · When the decline actually started

It wasn't five years ago. It was eight.

Year-over-year change in median view count. The big drops happened in 2018 and 2019 — coming off the 2017 cycle top. By 2020 the median was under 800. The 2021-onward period is a shallow grind, not a cliff.

Year-over-year change in median views

Negative bars only. The deeper the bar, the bigger the drop from the prior year.
SOURCE: derived from yearly medians, wcn-podcast/data.json snapshot
YearnMedianMeanMax

Pre-2021 (216 eps): median 2,015 · mean 3,587

Post-2021 (266 eps): median 510 · mean 541

Median change pre→post 2021: −74.7%

Reading: If you split the show into pre-2021 and post-2021 halves, the post era is a quarter of the pre era by median views. But most of that gap was already locked in by 2019, before the split point. The "five-year window" intuition under-counts the decline by about three years.
Section IV · The remaining audience

Smaller, more engaged — by the only measure on disk.

Watch-through (average view duration) requires YouTube Analytics access and is not present in the yt-dlp metadata. The closest on-disk proxy is the like-to-view ratio. Sample is limited — only 45 episodes have full .info.json sidecars with like and comment counts — but the directional signal is consistent.

Median like rate (likes / views), by year

A higher line means viewers were more likely to react. The post-2018 era runs ~1.5–2× the 2014 baseline.
SOURCE: 45 episodes with on-disk .info.json metadata · sample is uneven across years, treat as suggestive not definitive
Important caveat: the post-2020 engagement-rate sample is dominated by re-uploads of early episodes to the WCN channel. Those re-uploads tend to be served only to viewers who actively searched — a self-selected fan audience — which would inflate the like rate independent of any show-quality change. The directional finding (smaller, more engaged audience) is consistent with the data, but the data does not prove watch-through has held up.
Section V · Production cadence

Output held. Then it rose.

Episodes per year. 2021–2024 averaged 54 episodes/year — more than the 2017 peak year (47). The show kept producing through the entire decline. This isn't a fadeout in output.

Episodes produced, per year

2013 (1 ep) and 2026 (9 eps through March) are partial years.
SOURCE: wcn-podcast/data.json canonical episode list
Section VI · The co-host question

Where did everyone go?

Thomas asked whether views collapsed when Jimmy Song and Tone Vays left TBG. Built from a regex pass over all 482 transcripts plus a 2026-05-17 yt-dlp pull of their solo channels. Same data-only rules as the rest of the report.

Tone Vays · last sustained TBG run

TBG-175 · 2018-03-10

Joined TBG-077 (2015-08-21). Ran nearly weekly for 91 episodes through March 2018. After that: 9 sporadic guest appearances across 8 years.

Jimmy Song · last sustained TBG run

TBG-178 · 2018-06-01

Joined TBG-134 (2017-03-24). 64 appearances mostly clustered in 2017–2018. Returned as occasional guest 2020 onward; never resumed weekly cadence.

Both departures landed in 2018, within three months of each other. So did the steepest leg of TBG's view decline.

Median view count, year by year — TBG vs. the three co-host channels

Each line is the median view count per video for that year. Orange vertical markers are the Jimmy and Tone TBG-departure dates.
SOURCE: TBG from wcn-podcast/data.json snapshot 2026-03-18; Tone/Jimmy/MadBitcoins channels from yt-dlp pull 2026-05-17 · samples are partial (see methodology)
Reading the chart: In 2017 (everyone still on TBG), Tone's solo channel ran at roughly half TBG's median (4,695 vs. 8,768). In 2019 — the first full year after both departures — Tone's solo channel had risen to 10,835 median while TBG had dropped to 1,413. The two lines crossed. Tone stayed in the 8K–9K range through 2021 before contracting with the rest of the bitcoin-podcast space. Jimmy's solo channel sample starts in 2021 and tracks at roughly the same scale as post-departure TBG — it doesn't show the same audience-capture pattern.

TIB · the assumption check (revised 2026-05-17, second pass)

Correction (1st pass): An earlier version of this section read TIB from the on-disk mirror only and found 7 episodes (2018-04-01 to 2018-05-14). That was an undercount. The bulk of TIB lives as unlisted uploads on the WCN channel, invisible to the channel's public /videos feed and therefore missed by the mirror script. Pulling the WCN Today in Bitcoin playlist directly via yt-dlp surfaces 266 episodes (261 still public/unlisted, 5 set to private), running from 2017-07-03 to 2021-02-03 — a 3.5-year daily-ish news show, not a six-week one.

Correction (2nd pass): The 1st-pass conclusion that Jimmy and Tone "were never on TIB to leave" was wrong — it was based on a title-only scan. TIB titles are news headlines, not credits; they don't name cohosts. A transcript scan of the recovered TIB episodes (Whisper-base output, automated cohost-intro pattern matching) shows that Jimmy and Tone were regular voice cohosts on TIB, slotted into recurring segments alongside Thomas's news monologue.

Update (3rd pass, 2026-05-17): the transcript sample is now 5.7× larger (102 vs. 18) and produces a cleaner year-over-year drop-off — see the per-year table below. The pattern is unambiguous: nearly half of 2017 TIB episodes name Jimmy and 40% name Tone Vays; by 2018 that falls to ~10–14%; from 2019 onward, zero episodes name either. Mention rate does not simply track production volume (2018 had 60 episodes and 28 are scanned — 4× the 2019 sample). The wider-sample number kills the open methodology question from the 2nd pass.

Wider-sample update (2026-05-17, mid-recovery): the original 18-transcript pass has now been re-run against 102 of 260 TIB transcripts (39%, up from 7%). The early-era skew is reduced and the per-year breakdown is now the lede, not a 7%-sample anecdote.

YearTIB epsScannedJimmy epsJimmy %Tone Vays epsTone %
2017161612845.9%2439.3%
20186028310.7%414.3%
201913500.0%00.0%
202050
202121800.0%00.0%

The 2017→2018→post-2018 drop-off is the smoking gun for "they left both shows at the same time." In 2017 (everyone still on TBG), nearly half of scanned TIB episodes mentioned Jimmy and 40% mentioned Tone. By 2018 those rates had collapsed to ~11–14%. From 2019 onward — across 13 scanned post-departure episodes spanning 2019, 2021 — zero mention either name. The 2020 row has no transcripts in the current wave (5 unrecovered episodes; data pending).

The cohort-intro phrasing is not ambiguous — it's the same recurring-segment language across episodes:

"We're also joined by Jimmy Song with a check of the scaling. How's it going, Jimmy?" — 2017-08-15
"We're joined by Jimmy Song for a check of Bitcoin scaling. How's it going, Jimmy?" — 2017-08-17
"We're joined by Tone Vays. How's it going, Tone? Oh, going good. Last day here at the conference…" — 2017-10-08
"We're joined by tone veys for a quick check of the markets. Tone, are you with us?" — 2017-09-15

These are structured rotating-cohost segments — Jimmy on scaling, Tone on markets — not one-off interview drops. So the corrected literal answer to "did they leave TIB too?" is yes: they were on TIB regularly enough that the show carried recognizable voice-cohost branding, and when they stopped, that segment structure went with them.

The view trajectory still tells the same story it did in the 1st-pass revision. TIB's per-episode median views by year (TBG-departure dates marked):

Split by the TBG cohort-departure dates: episodes from launch through Tone's last sustained TBG run (2017-07-03 → 2018-03-10) had 185 eps at median 8,400; the in-between period (Tone gone, Jimmy still on TBG, 2018-03-11 → 2018-06-01) had 31 eps at median 4,000; everything after both were gone from TBG (2018-06-02 → 2021-02-03) had 44 eps at median 462. ~18× collapse in TIB median views across the TBG-cohort-departure window — directionally the same shape as TBG's own decline (43× peak-to-trough).

TIB · monthly episode count + median views (2017-07 → 2021-02)

Bars: episodes uploaded that month. Line: median view count, log scale. Vertical guides at Tone Vays' last sustained TBG run (2018-03-10) and Jimmy Song's (2018-06-01). The June 2018 → September 2018 gap is a three-month publishing pause that lines up with Jimmy's TBG departure within weeks.
SOURCE: WCN "Today in Bitcoin" playlist (260 episodes with dates), yt-dlp pull 2026-05-17

So the revised reading: the WCN-brand-fatigue explanation and the cohort-departure explanation are not in opposition; the transcript evidence shows they describe the same event. Jimmy and Tone's TIB segments stopped concurrent with their TBG departures, and TIB lost both its panel-format anchor and the audience pull those names carried. The fact that titles never credited them masked the change. The most defensible reading is now: Jimmy and Tone were on TIB as recurring voice cohosts, they stopped being on TIB at the same time they left TBG, and TIB's collapse trails that departure on the same shape and timeline as TBG's own. Whether bear-market sentiment and brand fatigue independently contributed is unresolvable from this data, but the "they left both shows" framing in Thomas's original question turns out to be substantively correct.

One concerning detail: TIB episodes from January–February 2021 have view counts in the 87–956 range, even though Bitcoin was at $30K–$40K and crypto news consumption was near an all-time peak. That's not bear-market suppression — it's audience-has-moved-on. The brand never came back, even when the asset did.

Methodology note (transcript scan): Transcripts are Whisper-base output (no diarization). Names matched with case-insensitive word-boundary regex; "Tone Vays" requires the literal pair (with common Whisper-garbled variants: veys, vase, vais, vaze, vayze). Bare-"tone" counts are filtered against false-positives (tonebase, tonality, monotone, undertone, "tone deaf", etc.) but not used in the per-year table — only the unambiguous "Tone Vays" pattern. The wider-sample re-run covers 102 of 260 TIB episodes (39%), up from 18 (7%) in the first pass; recovery for the remaining 158 is still in flight (round-1 PID 18297, in progress 2h44m at time of writing, 93/258 complete). The current sample is no longer early-era-skewed — every year 2017–2021 has scanned coverage. The 2020 row shows zero scanned episodes because all 5 TIB episodes from 2020 happen to fall later in the recovery queue; expect that gap to fill on the next pass.

Does the data support "views followed the co-hosts out"?

For Tone Vays — partially yes. His channel's median rose 2.3× from 2017 to 2019 (4,695 → 10,835) the same window TBG fell 6.2× (8,768 → 1,413). Some of TBG's lost audience appears to have moved to his channel. Treat this as evidence, not proof — the 2018–2019 bear market and broader attention shift are confounded with the departures, and we don't have the counterfactual.

For Jimmy Song — no clear pattern. The on-disk sample of his channel starts later (2021) and runs at the 200–500-median scale. He did not visibly capture a transplanted TBG audience.

The whole space contracted. 2023–2026 medians are lower than 2019–2021 medians on every channel in the comparison — Tone, Jimmy, TBG. The post-cycle leg of TBG's decline is consistent with industry-wide attention drift, not with the 2018 departures.

Section VII · The microphone direction

Who gets interviewed.

A separate question Thomas asked: how often is each of the three interviewed by other channels, versus how often Thomas interviews people on his own shows? Built from yt-dlp searches and a title-pattern scan over 1,934 WCN + MadBitcoins episodes.

Jimmy Song — guest appearances on other channels
146
Top hosts: Bitcoin Magazine (11), Tone Vays (9), MadBitcoins (5)
Tone Vays — guest appearances on other channels
86
Top hosts: WCN (5), AIBC World (3), Crypto Banter (2)
Thomas Hunt — guest appearances on other channels
25
Top hosts: Tatiana Moroz (3), BitcoinInvestor (2), Ugly Old Goat (2)
Thomas's own interview-format episodes
91
Proof of Work + Interview with X · peaked 2019 (48 eps)

Thomas as host: interview-format episodes per year

Proof of Work + "Interview with X" episodes on the MadBitcoins and WCN channels. Excludes TBG and TIB.
SOURCE: title-pattern scan, 1,934 MB+WCN .info.json sidecars
Reading the asymmetry: Jimmy gets interviewed roughly 6× more often than Thomas in the searchable data; Tone roughly 3.4×. Some of this is structural — long-running show hosts hold the mic, they don't hand it over often. Thomas's own interview-host activity peaked in 2019 at 48 episodes (mostly Proof of Work + carryover from Baltic Honeybadger 2018 conference coverage), then dropped sharply. He has hosted very few interview-format episodes since 2020.

One detail in the cross-pollination: Tone's top hosting channel in the search is World Crypto Network (Thomas's own channel) and Jimmy's #2 hosting channel is Tone Vays. Their interview maps are not three independent careers; they all interviewed each other before the departures.