My book review is on Mustafa Suleyman’s new book, The Coming Wave: Technology, Power, and the Twenty-first Century’s Greatest Dilemma. Suleyman is the founder of Deep Mind and has recently gone to work at Microsoft in an acquihire of his new company, Inflection AI. Surprisingly, the book is quite readable. In it, he unpacks the risks of AI and explores whether containment is possible. Takeaways include:
- We are at a point in history where one smart and motivated person could kill a billion people. (p. 27) What are we going to do about that?
- According to Suleyman, “containment encompasses regulation, better technical safety, new governance and ownership models, and new modes of accountability and transparency, all as necessary (but not sufficient) precursors to safer technology.” (p. 55) One of my biggest surprises of the book is that he doesn’t delve into blockchain or Web3 technologies at all. This seems to me to be one viable route to changing “governance and ownership models.”
- Up until now, our world has enjoyed waves of new technologies, resulting in population growth and improved standards of living. Things have changed. We are now faced with having to contain technologies that might potentially be unleashed on us without our consent. Said differently, “For most of history, the challenge of technology lay in creating and unleashing its power. That has now flipped: the challenge of technology today is about containing its unleashed power, ensuring it continues to serve us and our planet.” (p. 67-68)
- We now can create completely new molecules where we are. Synthetic biology is “the ability to read, edit, and now write the code of life….new techniques like enzymatic synthesis are faster and even more efficient while being less prone to failure….This is the promise of evolution by design, tens of millions of years of history compressed and short-circuited by directed intervention…. It brings together biotechnology, molecular biology, and genetics with the power of computational design tools.” (p. 110-111)
- As quantum computing moves from development into the early stages of reality, we now face a situation where cryptography, which is increasingly used to protect online identities, is at risk.
- The coming wave of technologies has four characteristics: the impact is largely asymmetrical; they are developing with incredible speed, can be used for many purposes, and are increasingly autonomous.
- Asymmetrical – while previous technologies have “created new threats, redistributed power, and removed barriers to entry,” never before have we seen this kind of asymmetry. A traffic accident was local but now interconnected cars can potentially wreak havoc. When Russia invaded Ukraine, a group of drone hobbyists went up against a well-trained military attack and were amazingly successful.
- Speed – Suleyman refers to the incredible speed as a hyper-evolution, with technologies iterating and evolving at a dizzying pace. Speed alone makes containment much more difficult. The question here is whether Moore’s Law holds in this new future.
- General purpose – These new technologies can be used in a host of ways, and often do not require large resources. According to Suleyman, “The next forty years will see both the world of atoms rendered into bits at new levels of complexity and fidelity and, crucially, the world of bits rendered back into tangible atoms with a speed and ease unthinkable until recently.” (p. 141)
- Autonomy – The risk that technology will take over what humans think and do is increasing as technologies become more autonomous. Already quantum computers are edging towards the point that humans cannot understand, or the technologies cannot explain what they are doing, sufficiently for human understanding.
- Much has been written by many about the risks of AI and the need to contain development before a catastrophe occurs. One challenge is capitalism. How can a “distributed, global, capitalist system of unbridled power” be persuaded to stop or curtail AI development? (p. 173)
- It’s easy to see how the global power structure may be shifted as more and more countries explore powerful AI research and development. Suleyman notes that the trend in scientific work is towards openness and collaboration. It makes me wonder even more about how we can better collaborate across the globe, as more and more people pursue this work. In looking at what the incentives are for those exploring this work, we cannot forget ego or greed.
- “Today’s cyberattacks are not the real threat; they are the canary in the coal mine of a new age of vulnerability and instability, degrading the nation-state’s role as the sole arbiter of security.” (p. 205-206) He writes about the challenges of balancing concerns about failed nation-states, authoritarian regimes that build massively controlling types of technologies, and the redistribution of global power that could come from extreme wealth, and/or powerful technologies. He notes that “[more] than one hundred U.S. towns have even acquired technology developed for use on the Uighurs in Xinjiang” as an example of a “textbook failure of containment.” (p. 245)
The final chapter details ten steps to containment. I’ll let you read the book if you’ve made it this far in my attempt to review a book on a complex subject.
