The diagnostic scores any information technology from 1 (power diffused) to 5 (power concentrated) on five dimensions, judged at the technology’s maturity point, when its social form has settled. The first four describe the channel: who owns it, what it costs to enter, who can use it, where the chokepoints are. The fifth, D5, is different. It asks where the capacity to originate new capability actually sits, with the user, or inside infrastructure the user can never own.
The five dimensions
- D1Ownership of the means of production. Many independent producers, or one firm and a cartel?
- D2Capital intensity / cost to participate. Cheap to become a producer, or rare and costly infrastructure?
- D3Access / reach of use. Effectively universal, or gated to a licensed few?
- D4Control of the channel / chokepoints. No controllable bottleneck, or a switch one actor can throw?
- D5Locus of generative capability. Does originating new capability rest with the user, or require infrastructure the user cannot own? Running or fine-tuning an existing model is use, not creation; it does not lower D5.
The coded lineage
| Technology | D1 | D2 | D3 | D4 | D5 |
|---|---|---|---|---|---|
| Printing press19th century | 2 | 3 | 3 | 3 | 2 |
| Telegraphlate 19th c. | 5 | 4 | 3 | 5 | 2 |
| Telephonemid-20th c. | 5 | 4 | 2 | 5 | 2 |
| Filmstudio system | 4 | 5 | 2 | 4 | 4 |
| Radionetwork era | 4 | 4 | 2 | 4 | 4 |
| Broadcast televisionnetwork era | 5 | 5 | 1 | 4 | 5 |
| Internetopen-web era | 2 | 2 | 2 | 2 | 1 |
| Social mediaplatform era | 4 | 1 | 1 | 4 | 2 |
| Frontier AI / computecurrent | 5 | 5 | 2 | 5 | 5 |
How to read it
Run your eye down the first four columns and you see Wu’s Cycle: the channel concentrates, then reopens, then concentrates again. The telegraph and telephone are near-monopolies on ownership and chokepoints; broadcast TV is heavy across the board; the open internet drops everything back toward diffuse. Nothing new there. That is the pattern history already taught us.
Now read the D5 column on its own.
For two centuries the capacity to create kept coming home to the user. Then it didn’t.
D5 sits low (2) through the telegraph and telephone, because those were common carriers: they moved your words, they never authored them. It rises with film and radio, spikes to 5 with broadcast television, then collapses back to 1 with the open internet, the high-water mark where anyone could originate and publish. Social media kept creation with users (2) even while gatekeeping the channel. And then frontier AI takes D5 to 5 again.
Two fives, opposite fates
Only two technologies in the lineage score a 5 on D5, and a critic reads that coincidence as proof there is nothing new. The opposite is true: they reach the same score by opposite mechanisms, which is why one reopened and the other may not.
| D5 = 5 | Mechanism | Did it reopen? |
|---|---|---|
| Broadcast TV | Reach-driven: spectrum and studio-scale distribution. Creating video was never the locked bottleneck. | Yes. Cable, then the internet, removed the reach barrier. |
| Frontier AI | Capability-driven: frontier compute, capital-bound and supply-chain-concentrated. Creation itself is the bottleneck. | Not guaranteed. There is no reach barrier to route around. |
That is the Inversion in one table: identical scores, opposite causes, opposite reversibility. The coincidence a skeptic reads as continuity is, on inspection, the clearest sign of the break.
Scores are reasoned judgments against anchored definitions, not measurements; maturity-point choice affects specific cells. The full rubric, per-cell rationale, and sourcing are in the working paper and its dataset (CC BY 4.0).