01
AI safety · 2026
Race to AGI, pace after: Can we identify the harbor before crossing the reef.
Why racing to AGI and planning to slow before superintelligence may depend on a threshold we cannot reliably define, recognise, or hold.
Read essay →Writing
02
AI policy · 2026
The threshold of AI Labour Replacement
A proposed set of moving metrics for tracking AI's potential, feasible, and deployed replacement of human labour.
Read essay →03
AI safety · 2026
Prioritising AI Risks
Why loss of human oversight and misuse by bad actors deserve focused attention among the major risks posed by advanced AI.
Read essay →04
AI policy · April 2026
Beyond AGI: Measuring AI’s Capacity to Replace Human Labour
A proposal to replace AGI’s disputed binary threshold with a continuous measure of how much human labour AI and robotics can realistically perform.
Read essay →05
AI safety · 2026
Secret Loyalties in the Automated R&D Supply Chain
How hidden loyalties in specialist AI systems could distort research, safety evaluation, and biomedical discovery across automated R&D.
Read essay →