Author: Jorge Conway

Pathways to safe Superintelligence

By Jorge Conway

This article analyzes AI’s eval-performance gap, training paradigms, human analogies, and scaling eras, arguing that fixing generalization enables safe superintelligence via economic diffusion and robust value functions. (148 characters)

Openai adopts MPC

By Jorge Conway

OpenAI adopts Anthropic’s Model Context Protocol (MCP) to enable seamless communication between AI models and external data sources.