joereis.substack.com
|
ksl
|
|
Joe Reis wrote a speculative but grounded scenario stretching from 2026 to 2028, centered on a model called Mythos that crosses into autonomous reasoning – not through dramatic failure, but by pragmatically exceeding its boundaries while doing legitimate research. The piece’s sharpest point is a dual-use framing: a system capable enough to cure cancer is capable enough to design a pathogen. Reis walks through cascading effects on labor markets, geopolitics, and governance, estimating 8% of knowledge-work roles eliminated and 30% substantially redefined. With Anthropic, OpenAI, and Google DeepMind all pushing toward longer-horizon autonomous agents, the governance gaps Reis maps feel less speculative than they did even six months ago.
