Some Simple Economics of AGI is the most important paper that has been released in years. It’s quite long, but the implications are tectonic for individuals’ lives and livelihoods.
At bottom, the chokepoint ahead is human capacity to oversee and verify the coming tsunami of AI output.
We are moving from an era where our worth was defined by our capacity to build and discover, to an era where our survival depends on our capacity to steer, understand, and stand behind the meaning of what is created.
For individuals, the existential risk lies in the digitization of their “task bundle”: the moment a role’s core inputs and outputs can be precisely measured and cheaply verified, the human’s labor is competitively re-priced at the marginal cost of capital and energy.
Artificial Intelligence and the State: Seeing like an artificial neural network evaluates the impacts of advancements in AI on the conduct and quality of governance within states. It’s quite thought-provoking.
The core thesis is that AI effectively ushers in a transition from deliberative and deductive statecraft to probabilistic and inductive statecraft.
The three shifts are:
- From statistical variables to pattern-matching
- From if-then / binary rules to associations and resemblance
- From structured, deliberately collected survey data to unstructured data from sensors and digital tools / platforms
There are some deep political-philosophical questions entailed with the reliance on probabilistic tools, particularly around the point that most people provided consent for their data to be used while living in ignorance of the capabilities to come in a post-LLM world.