How the watching path was made
This is the working behind the path, kept in the open: what I searched for, what came back, what I followed, how the list was sorted and what I kept. You can check it, rerun it, or do the same for a subject of your own.
Second-order cybernetics puts the observer inside what is observed. Here the observer is the curator: the searches are my choices, and so is every mark. Each pass changes the shelf by some amount, and the shelf is what those changes add up to. When a method stops changing it, it has run dry, and it is time for a different method, not more of the same one.
The steps
- Search. Each shelf has its own searches, run on YouTube and sorted two ways: most viewed and relevance. They are all listed below, each a link that runs it again.
- Follow. When a series or a channel kept turning up, I followed it whole, and filtered most of them down to talks on the subject. The list.
- Sort. most viewed (3 minutes or longer), and gems: like rate x comment rate x length, discounted by reach; under 50 views or 8 minutes is not a gem. A sorting aid, not a judgement.
- Review. A printed sheet per shelf; each talk ticked keep, maybe or skip by hand. Kept talks go on a public playlist, in the order I would walk them.
- Walk. The path, with a box to tick for each talk. What I have watched, I say so there.
From found to kept
Each shelf’s page lists every talk its searches found, kept ones first. On every shelf, following series and channels turned up more than searching did.
Each row is a shelf: what its searches and follows found, what made the printed review sheet, and what I kept. The scale is logarithmic, because the narrowing is that steep. Hover a mark for its number; all of them are in the table.
| Shelf | Found | By search | By following | On the sheet | Kept | To watch | Skip | Watched |
|---|---|---|---|---|---|---|---|---|
| Founders | 677 | 418 | 430 | 130 | 3 | 4 | 0 | 1 |
| Second order | 804 | 292 | 676 | 151 | 2 | 25 | 0 | 0 |
| Management and the VSM | 482 | 155 | 384 | 91 | 3 | 0 | 0 | 0 |
| Today | 722 | 152 | 618 | 100 | 2 | 15 | 0 | 0 |
| Systems thinking and dynamics | 711 | 153 | 632 | 81 | 5 | 0 | 0 | 1 |
| Complexity | 537 | 188 | 411 | 80 | 0 | 0 | 0 | 0 |
| Control and feedback (the engineering side) | 535 | 118 | 459 | 80 | 0 | 0 | 0 | 0 |
| Around it · Learning across fields | 1 | 0 | 1 | 0 | 1 | 0 | 0 | 1 |
| All shelves | 4,469 | 1,476 | 3,611 | 713 | 16 | 44 | 0 | 3 |
Observed October 25, 2026. A talk can be found both ways, so the two middle columns add up to more than found. A kept talk is counted under the shelf whose searches found it; on the path it is filed wherever it fits. Watched counts only what I have marked watched.
How the talks were found
Search finds what a platform thinks a subject is. Following finds what people who care about it have gathered. Most of this shelf came the second way.
Each bar is a shelf’s talks as shares: found only by search, by both, or only by following a series or channel. The figure on the right is the share only following found.
The numbers
| Shelf | Only by search | Both ways | Only by following | Only by following, share |
|---|---|---|---|---|
| All shelves | 858 | 618 | 2,993 | 67% |
| Founders | 247 | 171 | 259 | 38% |
| Second order | 128 | 164 | 512 | 64% |
| Management and the VSM | 98 | 57 | 327 | 68% |
| Today | 104 | 48 | 570 | 79% |
| Systems thinking and dynamics | 79 | 74 | 558 | 78% |
| Complexity | 126 | 62 | 349 | 65% |
| Control and feedback (the engineering side) | 76 | 42 | 417 | 78% |
| Around it · Learning across fields | 0 | 0 | 1 | 100% |
Over time
A dated line each time the shelf changes: how much each pass moved it. So far it is all one day, a point rather than a slope; the slope comes with time.
| Pass | Date | Found | On the sheet | Kept | Watched |
|---|---|---|---|---|---|
| 1 | 2026-10-25 | 2,933 | 280 | 0 | 0 |
| 2 | 2026-10-25 | 4,424 | 561 | 11 | 1 |
| 3 | 2026-10-25 | 4,424 | 668 | 14 | 1 |
| 4 | 2026-10-25 | 4,469 | 713 | 14 | 1 |
| 5 | 2026-10-25 | 4,469 | 713 | 15 | 2 |
| 6 | 2026-10-25 | 4,469 | 713 | 16 | 3 |
The searches
Run on October 25, 2026, each sorted most viewed and relevance. Each one is a link to the same search on YouTube today, which will not return exactly what it returned then.
Founders 11 searches
Second order 10 searches
Management and the VSM 7 searches
Today 8 searches
Systems thinking and dynamics 7 searches
Complexity 6 searches
Control and feedback (the engineering side) 5 searches
Around it · Learning across fields 4 searches
Followed whole
The series and channels I followed from start to finish, with the shelf each one fed. Filtered means only the talks on the subject came along.
Founders 48 sources
Second order 95 sources
Management and the VSM 52 sources
Today 52 sources
Systems thinking and dynamics 84 sources
Complexity 64 sources
Control and feedback (the engineering side) 41 sources
The rules the marks follow
- An empty mark means not reviewed yet, never rejected.
- A skip is a mark about fit for this path, not a verdict on anyone’s work.
- Every talk is credited and links to the creator’s own video. Nothing is copied or embedded.
- Counts are what was observed on the date given. Watched, and anything I say about a talk, is my word, shown only where I have said it.
- The ranking is a sorting aid, printed above so anyone can disagree with it.
Do it for a subject of your own
- Write the searches down before you run them, a handful for each corner of the subject.
- Run each one sorted by views and by relevance, and keep everything that comes back.
- Notice which series and channels keep turning up, and follow them whole.
- Print a sheet: the most viewed, and the well loved ones few people have found.
- Mark it by hand, and keep the counts each time you do, so you can see when a method has run dry.
If you make one, I would like to see it: hello@joshuaayson.com.
Everything found, shelf by shelf: Founders · Second order · Management and the VSM · Today · Systems thinking and dynamics · Complexity · Control and feedback (the engineering side) · Around it · Learning across fields
Back to the watching path · the reading path · domain cartography, the method this is a small case of