The author argues that the greatest risk from AI is not sentient machines but the socio-economic pressure to treat humans like machines. This trend, which predates AI, will be accelerated, devaluing human judgment and creativity in favor of standardized, replaceable labor, potentially leading to enslavement in 'human content farms'.
Drawing a parallel to the age of mechanical reproduction, the author posits that a flood of AI-generated content ('slop') will not destroy art. Instead, it will cause a difficult economic transition for some creators while ultimately creating a premium for bespoke, human-made work and enabling new artistic forms once artists gain precise, craft-like control over AI tools.
The author distinguishes between the intelligence of LLMs, which can manipulate symbols effectively, and their lack of wisdom. Drawing on Daoist philosophy, he argues that wisdom comes from experience beyond language, while LLMs are confined to the 'shadows' of reality represented by text, unable to access the deeper truths that language cannot capture.
Technology is presented not as an external force but as the most fundamental expression of humanity—the process of externalizing our minds and substantiating mental constructs in the world. We are defined by this co-evolution with our creations, from writing to AI, making the study of technology essential to understanding ourselves.
The modern surveillance society was not imposed by a totalitarian state as in '1984', but voluntarily constructed by consumers trading privacy for convenience. This dynamic sets a precedent for how future technologies, like agentic AI, will be adopted, raising questions about whether they will serve users or further entrench corporate data monopolies.
Keep pulling the thread on Ken Liu.