Breakout Signal

Channel profile · tracked by Breakout Signal

AI Discovery

21,500 subscribers · about 7 months old · ~3.1K subs/month to date ·publishing in English / Global ·Documentary / History, Feature Films & Series, Explainer / Science

4tracked breakouts
1.8×best breakout score
22.5×best reach, views per subscriber
749.3Ktotal views tracked
4.7×median reach, tracked videos

Breakout videos

  1. 01

    Grok AI Finally Reveals What's Hidden in The 1967 Patterson-Gimlin Bigfoot Film

    📖 Documentary / History ·483.9K views · 2.7K/day ·22.5× its audience · 26:03 ·2026-03-28 · English / Global

    1.8×
  2. 02

    A Texas Hog Eradication Operation Was Filmed — and It Caught Something Disturbing

    🎬 Feature Films & Series ·92.8K views · 1.2K/day ·4.3× its audience · 27:15 ·2026-07-07 · English / Global

    0.4×
  3. 03

    Loch Ness Monster Is Real... And It's Far Bigger Than You've Ever Seen

    🔬 Explainer / Science ·107.6K views · 4.1K/day ·5.0× its audience · 34:06 ·2026-08-30 · English / Global

    0.3×
  4. 04

    Grok AI Was Asked If Patterson-Gimlin Film Was Real — The Answer Shocked Experts

    🎬 Feature Films & Series ·64.9K views · 822/day ·3.0× its audience · 28:58 ·2026-07-08 · English / Global

    0.3×

How to read this channel

AI Discovery is a 21,500-subscriber channel (about 7 months old) working mainly in Documentary / History, Feature Films & Series, Explainer / Science, publishing first in English / Global. The radar tracks 4 of its uploads. Its strongest signal (breakout score 1.8× the dataset median) reaches 22.5× the channel's subscriber count and has averaged 2,674 views a day since it was published 181 days ago, on “Grok AI Finally Reveals What's Hidden in The 1967 Patterson-Gimlin Bigfoot Film” (483,910 views). That is a lifetime average, not a current rate — one capture per video means this dataset cannot distinguish a video that is still climbing from one that already peaked.

There is one statistic this page deliberately does not print: a per-channel “share of uploads that reached 3×”. Every video on this site cleared that floor before it could be tracked at all, so the figure would read 100% for every channel and would distinguish nothing. What is left instead: 4 tracked uploads pulling 749,250 views between them, the strongest reaching 22.5× its audience. Repeated strong reach across several uploads is the part that suggests a format; one outlier is not. The hook patterns the tracker infers from its titles are authority claim hook (2), statement / cold open (1), curiosity gap hook (1), and the production choices that repeat are no-face format (no host on camera), ai-generated visuals declared in the title, single-sitting long-form (30min+), ends on a cliffhanger to farm next-episode views.

What to copy and what to skip: copy the packaging pattern (title structure, the promise made in the first seconds, the episode length that keeps working) and skip the parts that only this channel can do — an existing subscriber base, cross-posting, or a back catalogue that feeds recommendations. Then compare this channel against the others on its category board: same niche, different market, different length. That comparison is where the transferable part usually hides.

Reproducible playbook

Opening hook (title-derived)

Authority claim hook · estimated hook window 18s before the main body.

Beat rhythm (Cold open → context → segmented build → reveal)

  • Hook 0%–4%
  • Context 4%–18%
  • Escalation 18%–55%
  • Peak 55%–82%
  • Close 82%–100%

Production choices that repeat

  • No-face format (no host on camera)
  • AI-generated visuals declared in the title
  • Single-sitting long-form (30min+)
  • Ends on a cliffhanger to farm next-episode views

Derived from the title, duration and market — no frame-level video analysis, and no use of the labels this site writes on an entry.

Compare with

Same niche, market by market

Niche board and neighbouring channels

Snapshot rebuilt 2026-09-25 · 2557 signals · 954 channels. Metrics come from the YouTube Data API v3 and are stored captures, not live counters; ages are recalculated from each video's publish date at build time.