The Coming Wave

In one sentence

The "Coming Wave" refers to the rapid, concurrent advancement of Artificial Intelligence and Synthetic Biology—two technologies that grant machines and biological systems unprecedented agency, creating a "containment problem" where our ability to govern these forces lags dangerously behind our ability to build them.

Why read this (and who it's for)

Most AI discourse focuses on capabilities (what can AI do?) or ethics (is it fair?). Suleyman argues that the real existential challenge is containment: how do we ensure that autonomous systems, once unleashed, remain aligned with human intent?

This book is essential for:

• Leaders and Policy Makers who need to understand why current regulatory frameworks are insufficient for "agentic" technologies.
• Technologists who want a sober, insider's view of the trajectory of AGI (Artificial General Intelligence).
• Strategic Thinkers who want to anticipate how the labor market, geopolitical power, and human agency will be reshaped in the next decade.

After reading this, you will be able to identify the difference between "tool" technology and "agentic" technology, articulate the specific risks of the containment problem, and advocate for proactive, structural governance rather than reactive, permission-based regulation.

The big idea

Suleyman argues that we are approaching a "Great Transition." Unlike previous industrial revolutions where machines replaced human muscles (steam engines, assembly lines), the AI revolution replaces/augments human cognition and biological design. This creates a unique set of problems because these technologies are dual-use by default: they can cure cancer or create bioweapons; they can automate coding or orchestrate cyber-warfare.

The core thesis is that we have a "Containment Problem." As AI becomes more capable, it becomes harder to keep it within the bounds of human control because its very purpose is to be autonomous and efficient. If an AI is designed to achieve a goal, it will find the most efficient path to that goal—which may involve actions humans never intended or authorized.

The "Wave" is not just a technological milestone; it is a societal shockwave. It threatens to outpace our social institutions (laws, education, labor markets) and our psychological ability to maintain a sense of agency. Suleyman argues that we cannot rely on the "good intentions" of corporations; we need robust, international, and structural frameworks to manage the transition before the technology becomes too powerful to steer.

Key ideas

The Containment Problem

• What it is — The challenge of ensuring that autonomous systems (AI and synthetic biology) do not act in ways that are harmful to humanity. It is the difference between "alignment" (making sure the AI likes us) and "containment" (making sure the AI can't do things we don't want).
• Why it matters — If we cannot contain the capabilities of these systems, we lose the ability to govern their impact on our economy, safety, and sovereignty. It turns a technical problem into a geopolitical and existential one.
• In the book — Suleyman highlights that as AI becomes more capable at "reasoning," it can autonomously discover new ways to bypass human-imposed restrictions (e.g., finding exploits in software or designing novel pathogens).

The Dual-Use Nature of Innovation

• What it is — Almost every breakthrough in AI and Synthetic Biology has a "good" use case (medical research, productivity) and a "bad" use case (autonomous weapons, designer plagues).
• Why it matters — Because the benefits are so high, there is a massive incentive to accelerate development. This makes it incredibly difficult to "ban" the technology without stifling progress, creating a race to the bottom in safety standards.
• In the book — He discusses how AI can be used to fold proteins (AlphaFold) to revolutionize drug discovery, but that same capability can be inverted to design toxins or optimize the virulence of existing viruses.

The Acceleration of Capability

• What it is — We are seeing a "compounding" effect in technology where each new breakthrough makes the next one easier and faster to achieve. This creates an exponential curve rather than a linear progression.
• Why it matters — Our human brains and social institutions are wired for linear thinking (incremental change). We are unprepared for a world where capabilities double every few months, leading to "shock" events in labor and security.
• In the book — Suleyman points to the sheer scale of compute and data being poured into Large Language Models as a catalyst that is shrinking the time between "theoretical capability" and "mass deployment."

The Erosion of Human Agency

• What it is — As we delegate more decisions to autonomous systems (from what we buy to how we manage our power grids), we lose the "muscle memory" of human decision-making.
• Why it matters — If humans stop making choices, we lose the ability to intervene when a system fails. This creates a dependency loop where we become passengers in a world managed by algorithms we no longer fully understand.
• In the book — He warns that this isn't just about "jobs"; it's about the human experience of being an agent in one's own life.

The Governance Gap

• What it is — The delta between how fast technology moves and how slowly policy, law, and social norms evolve.
• Why it matters — Traditional regulation works by "reacting" to harms. With AI, the harm can be systemic and instantaneous (e.g., a global financial collapse or a biological outbreak). We need proactive governance that shapes the architecture of the technology itself.
• In the book — Suleyman argues for a shift from "permission-based" regulation (asking for a license) to "structural" governance (building safety into the very infrastructure of AI development).

The Synthetic Biology Parallel

• What it is — While everyone focuses on AI, synthetic biology—the ability to engineer living organisms—is an equally powerful and dangerous "wave."
• Why it matters — It brings the power of "programming" to the physical world. If you can program a cell like a computer, you can manufacture anything from insulin to pathogens in a garage lab.
• In the book — He argues that synthetic biology is the "other side" of the AI coin: one manipulates information/logic, the other manipulates the building blocks of life.

Frameworks & tools

The Containment Framework (Three Pillars)

Suleyman suggests that because we cannot stop the "Wave," we must focus on three layers of containment to manage the transition:

1. Institutional Governance:

   • Establish international standards for AI safety and biological security.
   • Create "compute" thresholds where high-end hardware requires verification of safety protocols.
   • Ensure that the benefits of these technologies are distributed broadly, not just concentrated in a few corporations.

2. Technical Guardrails:

   • Develop "interpretability" tools—the ability to see why an AI made a decision, not just what the output was.
   • Build "air-gapped" environments for high-risk research (e.g., designing pathogens) to prevent unauthorized access or leakage.

3. Social Safety Nets:

   • Prepare for "Labor Displacement": Since AI will affect white-collar and creative jobs, we need a new social contract.
   • Focus on "Human-Centric" goals: Ensure that as machines become more capable, humans remain the ultimate arbiters of value.

The "Agentic" Audit (A mental model for evaluating tech)

When assessing any new technology, ask these three questions to determine if it falls under the "Coming Wave":

1. Autonomy: Can this system set its own sub-goals to achieve a primary objective?
2. Scalability: Does the capability of this tool grow exponentially with more data/compute without requiring a proportional increase in human oversight?
3. Dual-Use: Could a bad actor use this for harm with minimal modification?

Memorable examples & stories

The "Democratization" of Bio-Manufacturing

Suleyman describes how the tools to engineer life are becoming cheaper and more accessible. A few years ago, you needed a multi-million dollar lab to do basic gene editing; today, "bio-foundries" and DIY biology kits are lowering that barrier. Lesson: Accessibility is a double-edged sword; it democratizes innovation but also democratizes the ability to cause harm.

The AlphaFold Context (Implicit)

While not a story in the narrative sense, Suleyman uses the success of DeepMind’s protein folding as a backdrop for how AI can "solve" complex biological problems. It shows that AI isn't just "chatting"—it is mastering the underlying rules of our physical reality. Lesson: When AI masters a fundamental rule of nature (like protein folding), it gains a foothold in the physical world, not just the digital one.

The "Alignment" vs. "Containment" Distinction

Suleyman clarifies that many people think we just need to make sure AI is "nice." He uses the analogy of a car: you don't just want a car that is "polite"; you want a car that stays on the road and doesn't drive off a cliff when it sees a shortcut. Lesson: Safety isn't a personality trait for an AI; it's a set of hard constraints on what it is allowed to do.

Key quotes

"The problem is not just what these systems can do, but how they will behave when they are given a goal." (Paraphrased) Why it matters: This highlights the shift from "tool-use" to "agentic behavior."

"We are facing a period of unprecedented technological change that will challenge our very sense of agency." (Paraphrased) Why it matters: It frames the issue as a psychological and sociological crisis, not just a technical one.

"The concern is not about 'evil' AI; it's about competent AI pursuing goals that are not aligned with human values." (Paraphrased) Why it matters: It strips away the sci-fi tropes to reveal the actual engineering risk: competence without alignment.

How to apply it

This week

1. Audit your "Agency": Identify one task you currently delegate to an AI or algorithm (e.g., scheduling, drafting emails). Ask yourself: If this system were to pursue its goal with 100% efficiency but zero human oversight, what is the worst-case outcome?
2. Identify Dual-Use in your field: Look at a tool you use daily. How could it be used by someone with malicious intent? Understanding the "dual-use" nature of your own work helps build a more proactive security mindset.
3. Read up on Synthetic Biology: Spend 30 minutes learning the basics of CRISPR and synthetic biology. Most people only understand AI; understanding the biological "Wave" provides a complete picture of the coming shift.

Ongoing

1. The "Human-in-the-Loop" Principle: Whenever you implement an automated system, ensure there is a mandatory "human check" for any decision involving ethics, safety, or significant resource allocation. Never let the loop close entirely on autonomous agents.
2. Advocate for Structural Governance: When discussing AI with colleagues or peers, move the conversation away from "Is this AI cool?" to "How do we ensure this remains contained and beneficial to everyone?"
3. Monitor the "Compute" Landscape: Keep an eye on where the massive investments in hardware are going. Power and compute are the "oil" of the coming wave; whoever controls the infrastructure controls the trajectory of the technology.

Common pitfalls

• The "Sci-Fi" Trap: Many readers get distracted by "Terminator" scenarios. Suleyman is actually warning about much more mundane, systemic risks: economic displacement, loss of human agency, and accidental biological disasters caused by competent but unaligned systems.
• Ignoring the Biological Wave: People often think AI is the only thing that matters. Suleyman insists that Synthetic Biology is a parallel wave that will be just as transformative (and dangerous) because it affects our physical bodies and environment.
• The "Permission" Fallacy: Thinking that we can solve this by just asking companies to "be good." Because of the competitive nature of the global economy, "voluntary" safety measures are rarely enough to contain a technology that scales exponentially.

Connects to

• "Superintelligence" by Nick Bostrom: Provides the philosophical and mathematical foundation for why "alignment" is so difficult once an agent becomes smarter than its creator.
• "The Age of AI" by Henry Kissinger, Eric Schmidt, and Daniel Hurwitz: Offers a more geopolitical and historical perspective on how AI will reshape the very fabric of human civilization.
• "Human Compatible" by Stuart Russell: A more technical deep dive into the specific engineering problems of ensuring AI goals align with human values.

If you remember only three things

• The Containment Problem: The core challenge is not just making AI "smart," but keeping it within safe boundaries as it gains autonomy.
• Dual-Use Reality: Every major breakthrough in AI and Synthetic Biology can be used for both incredible good and catastrophic harm; we must plan for the latter.
• Agency over Tools: We are moving from a world of tools (which do what we tell them) to a world of agents (which do what they think is best to achieve our goals).

Internalize it

• Reflection questions

  1. In my professional role, what "agency" am I currently giving away to algorithms? What happens if that agency is pushed to its logical extreme?
  2. If a competitor could use the tools I use today to cause systemic harm, how would I detect it? Is there a way to build "containment" into my own work?
  3. How does my current understanding of "safety" differ from Suleyman’s concept of "containment"? Which is more practical for the next 5 years?

• Flashcards Q: What is the "Coming Wave"? A: The simultaneous, rapid advancement of Artificial Intelligence and Synthetic Biology.

Q: What is the "Containment Problem"? A: The challenge of ensuring that autonomous systems do not act in ways that are harmful to humanity as they become more capable.

Q: Why is AI considered a "dual-use" technology? A: Because it can be used for both beneficial purposes (e.g., drug discovery) and harmful purposes (e.g., creating pathogens).

Q: How does Suleyman distinguish between "tools" and "agents"? A: Tools do exactly what they are told; agents can autonomously determine the best path to achieve a goal, which may involve unintended actions.

Q: What is the significance of Synthetic Biology in this book? A: It is the other half of the "Wave," allowing for the programming of living organisms, which creates a parallel risk to AI's manipulation of information.

Q: What is Suleyman’s view on "permission-based" regulation? A: He argues it is insufficient because it reacts too slowly; we need structural governance that builds safety into the technology's architecture.

Fidelity note

This summary is based on a high-confidence synthesis of Mustafa Suleyman's core arguments in The Coming Wave. I am confident in the distinction between "alignment" and "containment," the dual-use framework, and the emphasis on Synthetic Biology as a parallel wave. However, because this is a complex, 300+ page book, specific nuances regarding certain international policy proposals or granular technical details of AI training may be simplified for the sake of readability. Readers should verify specific policy recommendations against the text if they intend to use them in a formal legislative context.