Agency
How tools, memory and persistence turn a model into a system that acts.
Artificial intelligence
The Consequence Machine
AI Agents, Superintelligence and the Fight to Stay in Control
AI is moving from generating answers to taking action. This book explores what that shift means for the people who build, deploy and live alongside it.
An ebook sample is available on Google Play Books.
About the book
When intelligence
can take action.
The Consequence Machine follows the development of AI from foundation models into systems with tools, memory and the ability to carry out work. It distinguishes intelligence from agency and autonomy, and examines how permission turns capability into action.
The book considers the commercial and geopolitical pressures behind deployment, the infrastructure that supports it, and possible paths towards more capable successors. Throughout, the focus returns to human control: who grants authority, how actions are checked and how a system can be interrupted or corrected.
Richard’s writing sits alongside his practical work at AI Venture X. His public engineering examples explore approval, bounded permissions and recovery in runnable reference workflows.
Consequences
The effects on work, institutions and the distribution of power.
Control
Permissions, oversight, verification and the ability to intervene.
Interviews & book enquiries
Continue the conversation.
For media, podcast and book enquiries, connect with Richard on LinkedIn.