Deconstruction of discourse: the vast majority of humans out of the decision loop.
All the main online sources are moving to subscription models with higher subscriptions for the real information. Only those with money will have access to information that informs decisions. At the low end, information is bundled only for its entertainment or propaganda value.
At the same time, decisions are increasingly automated as vast data streams are digested by automated processes.
20 or 100 years ago, people could stop work or stop buying or protest in the street. 20-100 years from now, there will be nothing the vast bulk of people can do to change their fate.
The resulting lack of citizen governance will at best be a world broken into geographic silos headed by corporate keiretsu.
So the next big technical thing will be domain-specific semantic models to drill down past what AI can do with probabilistic models -- just as the economy has moved well past bulk goods to bespoke services, e.g., the Nature article today providing a model for immune system cell-surface-protein interactions: https://www.nature.com/articles/s41586-022-05028-x
I think it's worse. Pre-internet, people paid for access to the same information, information written by journalists, distributed by a select number of papers.
Today, and perhaps worse going forward, this is reversed: a select number of persons pay for access to journalistic quality. The rest consume what the market hands to them, for whatever underlying reason.
If you're correct, the current intermediate condition where people have removed themselves from the (true) information loop but remain in the decision loop is the scariest state in that decline.
All the main online sources are moving to subscription models with higher subscriptions for the real information. Only those with money will have access to information that informs decisions. At the low end, information is bundled only for its entertainment or propaganda value.
At the same time, decisions are increasingly automated as vast data streams are digested by automated processes.
20 or 100 years ago, people could stop work or stop buying or protest in the street. 20-100 years from now, there will be nothing the vast bulk of people can do to change their fate.
The resulting lack of citizen governance will at best be a world broken into geographic silos headed by corporate keiretsu.
So the next big technical thing will be domain-specific semantic models to drill down past what AI can do with probabilistic models -- just as the economy has moved well past bulk goods to bespoke services, e.g., the Nature article today providing a model for immune system cell-surface-protein interactions: https://www.nature.com/articles/s41586-022-05028-x