fencing the non rivalrous data commons, and the sovereign internet that never really was

Lovely idea, the declarations of independence of cyberspace notwithstanding, the vision of a new politics is not standing up well to the greatfirewalls and random total isolation of states (yes, old fashioned states, as defined by geographic boundaries, including space and maritime, like china, iran, north korea, but now lots of little old places in Europe)....it isn't that hard to find all the ingress/egress points (including satellite updownlinks etc) and plug them after all...

Meanwhile, the data commons that was the World Wide Web is being rivalled to death bu hyperscalers, starting way back with google (search and click through) but now with AI.

Economicists love talk about data as non-rivalrous, because of their naive model of zero copy cost of bits, so anyone can "take" a copy, but the source is also left behind so there's no "loss" of value to the source.

Howver, attention (via meta data, provenance, attribution, even accounting/payment)....is very rivalrous - i only have two eyeballs, and one brain and so many hours each day, so it matters which copy I look at.

When the hyperscalers "add value" to the data, they subtract it from the originators. Worse, by trying to multiply the value (by combining), without recompense to the content creators, they undermine the motives of people to contribute any content. So the consequence is the death of Metcalfe's law (the value of the network is the square of the number of nodes -or at least is super-linear in the number of nodes, as any node that is added is both a consumer and a producer so the adding one more node to n, adds n+1 to the net value, not just 1). This means the hyperscalers' long term business plan is doubly dead. Their value cannot actually multiple once they trained on the common crawl - it will increase (at best) sub-linearly. In fact, as it de-motivates people from contributing any new content, the hyperscalers value will fall. We don't want to be data serfs to the AI overlords.

So AI wasn't an existential threat to humans, but it is an existential threat to knowledge.