AI as a subject, never AI as a method
The single largest competitor breakout in the window is a Top 10 about AI saying things scientists cannot explain, at 28x its channel median. WatchMojo has not made the list yet.
What was measured, and what it means here
Left, the four scores as computed for this subject. Right, a person's reading of each one for WatchMojo in particular.
- Strongest signal
- Demand at 95
- Weakest signal
- Fit at 17
- Videos on the subject
- 28, 11 of them owned
- Comments naming it
- 85, 70 unprompted
Second-highest demand score in the corpus: 85 comments, 70 of them under videos about something else. People bring AI up everywhere.
Upload share up 1.4x. Steady rather than spiking.
One competitor publishing on it: MostAmazingTop10, whose two AI lists ran at 28.1x and 7.9x and whose Part 2 ran at 3.2x. The sequel demand is visible in their own numbering.
The lowest fit score on this board, and it is honest: WatchMojo's videos tagged AI in the window are Feedback Friday episodes and tech-leader lists at 0.4x. MsMojo's Modern Tech That Already Doesn't Exist ran at 6.4x, which is the nearest real read.
Scores are 0 to 100 percentile-style ranks among the 262 subjects in this run, never absolute claims: a high demand score means most other subjects drew less conversation, not that most viewers want it. Fit is read from 11 owned uploads on the subject itself. Everything on the left is counted from the corpus; everything on the right was written by a person and cites its evidence further down the page.
When to publish
The call and the measurement side by side, so an override is visible rather than silent.
The competitor's Part 2 landed; the subject is proven and the sequel demand is measurable. No release beat to wait for.
The measurement alone would say: Evergreen. The author overrode it, for the reason given beside this.
- Momentum
- 55
- of 100, rank in this run
- Recent mentions
- 45
- comments in the last 14 days, 16 in the 60 before
- Late share
- 47%
- of comments arriving more than a month after upload
The rule is simple on purpose: hot conversation plus rising uploads reads as now; rising with less heat reads as within days; a subject still drawing comments a month after upload reads as evergreen; the rest waits for its moment in the cycle.
The shape the evidence points to
In the corpus's own vocabulary of ten format lanes, matched on titles.
The breakout is a Top 10 of moments. Number it 20 and keep the entries concrete and sourced.
Index is views over each channel’s own median for the same video type, on uploads at least three days old. An index above one means the format beats the channel’s typical upload; the two bars share a scale.
Where the editorial team would start
These are title directions for the editorial team, not scripts. Each is written in the shape the channel's own record rewards; the writing stays with the writers.
- 01Top 20 Times AI Did Something Nobody Programmed
- 0220 AI Predictions That Already Came True
- 03Top 20 Movies That Got AI Right Decades Early
What could make this call wrong
The audience's complaints about AI voiceover and AI text-to-speech are in this corpus too, and the May incident was about AI music in a tribute. Content about AI is welcome; content made by AI is the sore spot, and the founder has said publicly that the company has not replaced staff with it.
Weakest measured signal: Fit at 17 of 100 among the 262 subjects in this run.
What the call was read from
First the comments and videos the author cites. Then the live record, pulled from the corpus at render time: what WatchMojo has already made on this subject, and what broke out for competitors.
Comments the call cites
2 citedSkynet goes self aware 2029 we have been warned
AI generated text to sppech never sounds good. ANyway, I did learn what Secret Wars is all about.
Videos the call cites
1 citedWatchMojo’s own record on this subject
11 owned uploads old enough to index, top 8 by indexCompetitor breakouts on this subject
top 5 by indexComments are the top-sorted sample YouTube returns for each video, capped at our own limit, never a video’s full comment count. The index on each video is its views over its own channel’s median for the same type, on uploads at least three days old, which is why a small channel’s breakout ranks above a large channel’s ordinary day. Competitor supply means the 15 competitor channels collected for this run, not YouTube as a whole, and everything here comes from a single collection on 2026-08-27.
5 of the 20 inputs in the brief
The brief lists twenty inputs and asks that they be combined, not presented. These are the ones this call rests on most, with how each was measured in this run.
Subject share of uploads and of comments in the last 14 days against the 60 days before, as a lift ratio.
Every video's views over its own channel's median for the same type, on videos at least three days old.
Supply counted across the 15-channel competitor set. A YouTube-wide search pass per subject was written but not run: YouTube began bot-checking the collector after 1,700 requests.
YouTube only. Reddit's keyless feeds were measured on an earlier COIL engagement and yielded too little to be worth it; TikTok and search need paid access.
329 videos in the corpus beat their own channel median by 2x or more, ranked by the multiple rather than by views.