About this ebook
An AI model can write an answer. Connect it to memory, tools, money, code and real accounts, and the answer can become an action. Give it a goal, permission and time, and it can become an agent.
Companies and governments are racing to build systems that can do more with less supervision. The prize could be extraordinary: cheaper expertise, faster scientific discovery, better medicine, personalised education, continuous cyber defence and new forms of prosperity. But the same capabilities could automate fraud, deepen surveillance, disrupt work, concentrate power and turn a bad decision into an irreversible consequence.
The next threshold is even more consequential. AI is beginning to assist parts of the research process used to build better AI. That is not yet open-ended recursive self-improvement. But if enough links in the loop close, today’s agents could help create more capable successors.
In The Consequence Machine, British engineer and AI founder Richard Russell takes you from the breakthroughs behind modern foundation models to AI agents, the global contest for chips and compute, competing definitions of AGI and the conditional routes towards superintelligence.
You will discover:
- How contemporary AI actually works, without requiring a technical background
- Why intelligence, agency, autonomy and consciousness are different
- How tools, memory, permissions and persistence turn models into systems that act
- Why corporations and nations may fear slowing down more than continuing the race
- What AGI, recursive self-improvement and superintelligence would really require
- Where meaningful human control could still be preserved
This is neither a promise of utopia nor a prediction of extinction. It separates observed harms from emerging risks, plausible pathways and speculation. It also confronts the possibility at the heart of the race: humanity may need powerful AI to defend against harmful AI—without creating a guardian we can no longer question, replace or control.
The future has not been decided. But the permissions are already being granted.
Discover more
About the author
Richard Russell is a British author, engineer and AI leader, and the founder of AI Venture X. His writing explores intelligence, power and human agency: how people retain responsibility and meaningful choice as machines and institutions become more capable.
His non-fiction includes The Consequence Machine: AI Agents, Superintelligence and the Fight to Stay in Control; The Hollow State: Why Britain Still Votes But No Longer Chooses; and The Scent of Thought: Dogs, Humans and the Future of the AI Mind. His forthcoming novel, The Signal Inside, is Book One of The Intelligence Threshold.
Alongside his writing, Richard designs and operates agentic AI systems through AI Venture X, drawing on more than two decades of engineering, technology leadership and enterprise programme delivery. His practical work focuses on useful automation, clear permissions, human oversight and accountable decisions.
His creative work also includes Night Harbour Records, bringing together five music acts across jazz, pop, country, rock and synthwave. He lives in the United Kingdom with his family and their German Shepherd, Max.
