Import AI 458: Making Sense of the Future and a Singularity Scenario
Anthropic
This issue of Import AI features a speech delivered at Oxford's HAI Lab and a fictional story about a positive singularity. The speech examines how rapid AI progress presents society with a choice: explore the future or retreat from the present. The author describes their interaction with AI from spell-checking to deep personal advice and creating repeatable skills based on their own content.
Issue 458 of Import AI consists of a long-form essay based on a speech Jack Clark gave in 2026 at the Human-Centered AI Lab (HAI Lab) of the Oxford Institute for Ethics in AI, in collaboration with the Cosmos Institute, and a fictional story about a positive singularity. In the first part, the author discusses the Epoch Capabilities Index (ECI) chart, which shows AI performance growth on over 40 benchmarks, and lists key achievements of recent years: passing the bar exam (March 2023), silver and gold medals at the International Mathematical Olympiad (July 2024 and July 2025), co-authoring new mathematical proofs (2025), and finding software vulnerabilities by the Claude Mythos system. The author emphasizes that such progress continues thanks to hundreds of billions of dollars in investment in computing power, and warns of existential risks, including the possibility of recursive self-improvement of AI within the next two years. In the second part, the author describes how AI permeates his personal life: from text proofreading (summer 2023) to parenting advice (January 2026) and deep personal conversations. He tells how he used his own Import AI newsletter archive to create a reproducible skill that allows AI to generate charts and analyze hundreds of scientific papers, giving him a sense of discovery. The third part discusses the difficult choices facing individuals and society.
Source: Import AI —
original