love boring running shoes for training

cut the middleman

reasoning model and in context reinforcement learning

general reasoning capability and long context model are reshaping ai theoretical landscape

unreasonable effectiveness of social scaffolding

to what ends?

2 operation mode

exploring how humans need to balance between AI-assisted productivity and active learning modes in the post-AGI world

formidable exponential

drawing parallels between AGI and Bitcoin as digital transformations of intelligence and capital, respectively

o1 is the spark of agi

reflecting on how personal agency, choice, and will remain uniquely human even as AGI surpasses us in most tasks

keep trying

contemplating the value of genuine human connection and direct experience in a world where AI becomes the ultimate mediator

go faster and further alone

exploring how AI empowers individuals to achieve more independently, challenging the traditional wisdom of collective progress

language interface

examining how different languages and mediums serve as interfaces for expressing human intent to AI systems

meditation

understanding meditation as a practice of conscious choice and effortless control rather than forced emptiness or denial

good and evil

a thought experiment on how AGI could blur traditional notions of good and evil, and its implications for meaning and human nature

humor

exploring humor as the ultimate Turing test, requiring both intelligence and grounded experience to truly understand and create