AGI is Here
A viral AI essay argues the latest models meet key AGI criteria, capable of novel reasoning and planning.
Deep Dive
AI researcher Gordon Seidoh Worley argues on LessWrong that Anthropic's Claude Opus 4.6 and OpenAI's GPT-5.3, released Feb 5, constitute a 'minimum viable AGI.' He claims they demonstrate novel reasoning, planning, and goal achievement, surpassing junior engineer capabilities in coding tasks. The central debate is whether their current limitations are due to the models or their 'harnesses,' suggesting they represent a threshold in AI generality.
Why It Matters
If accurate, this shifts the professional conversation from building AGI to managing and applying systems with human-like general reasoning.