Essays on what AI makes possible, who benefits and which consequences we underestimate.
Read it. Listen to it. Let your agent investigate it.
When a model, tools and people produce a result together, how much can the system repeat?
OpenAI released 719 mathematical manuscripts from roughly 4,000 posed problems, with exceptions and human edits noted in the catalogue. Before a result counts as autonomous capability, ask what the specified system can repeat.
Who shapes the choices before the scientist decides?
Claude narrowed 3,500 candidate systems to 20 reports. The human gets the final decision. The AI has already shaped the options.
Can safety learning keep up with model development?
OpenAI stopped GPT-6.1 Astra's release. The harder deadline comes when an agent's work starts shaping its successor.
New essays by RSS, for you or your agent.
