The Coming Wave

For all of history, two things about technology have been true. Technology always gets cheaper and easier to use, which allows for widespread adoption, and the challenge has always been creating new technology. The latter claim is no longer wholly true. The challenge of “The Coming Wave” is containing technology.

The coming wave refers to upcoming AI and biotech advancements that will, according to Mustafa Suleyman, augment society in a way never seen before. A greater shift than those caused by language, the printing press, or even the steam engine. Before delving into what this wave entails and why/how it should be contained, it would be useful to lay out our current situation. A situation that will have already evolved by the time I am done writing this report.

AI

I bought this book for AI. Mustafa Suleyman is, after all, a co-founder and former head of Deep Mind (Google’s AI Company) and co-founder of Inflection AI. It wasn’t very long ago that, for most people, artificial intelligence was thought to be decades away and meant personal digital assistants akin to JARVIS from Iron Man. The AI that did exist was rudimentary and had very narrow use cases. The average person wasn’t having an existential crisis when they found out they could play chess with a bot. Suleyman uses a John McCarthy quote to capture this attitude. “As soon as it works, no one calls it AI anymore.” AI was, “what computers can’t do.” This view has been completely shattered since the debut of ChatGPT.

ChatGPT put artificial intelligence on everyone’s radar. OpenAI taught a machine to give almost humanlike responses to text prompts. Bard/Gemini, Rufus, Copilot, DALL-E, Firefly, and the list goes on. These AI models are what will lead to more advanced AIs defined as artificial capable intelligence (ACI). These models will be able to perform a range of complex tasks instead of being limited to just pictures or text. Microsoft Copilot will soon be able to give you presentation advice after looking through your company’s financial reports and evaluating your team members' performance. After ACI comes artificial general intelligence (AGI) which will be able to perform all human cognitive functions better than humans. That’s where you get JARVIS… and HAL (if we’re not careful).

Biotech

If biotech was a YouTube video, there’d be a banner ad on the screen saying, “Powered by AI.” Sequencing and synthesizing DNA has gotten exponentially faster and easier with AI. The field is getting frighteningly easier and easier to enter. For the first time, biology is growing at a speed previously reserved for technology. The Kilobaser DNA Synthesizer costs $35,500 and needs no kind of license to purchase. For an extra $14,000 you can synthesize modified DNA and RNA. Those motivated enough can currently create strands of diseases in their garage. “Unlimited Capabilities.” Ask AI to generate a cure for a virus… or a virus that doesn’t have a cure. Technology is evolving from controlling bits to controlling atoms.

How long until you can print corn in your kitchen, or have your medicine cabinet produce a vaccine for this season's cold? When will making chemical weapons become easier than grocery shopping?

The Coming Wave

There are four features of the coming wave. Asymmetry, endless acceleration, omni-use, and autonomy. Each is equally discomforting. I was unable to imagine a life in this world for more than seconds at a time. My brain rejected the concept like I was imagining being dead.

There is an asymmetry of power in the coming wave. With where AI and biotech are heading, each person will have more power than ever before. Power to do good or to do bad. The way billionaires still use the same smartphones as us, we will be using the same AI they have. There is also asymmetry in the way that the smallest of errors can produce a huge amount of havoc. If autopilot fails on a plane, one plane will crash. If an AI network that flies planes fails, all planes crash. Imagine all self-driving cars hallucinating empty roads and accelerating during gridlock or a defense system deciding to eliminate all children because they could grow up to become terrorists.

Endless acceleration means that society will not be able to catch up to the advancements in biology and technology. When the car was released, we had to develop roads, traffic rules, and a plethora of other systems because of automobile-related hazards. This AI acceleration won’t give us the luxury of time and we may not be able to afford any mistakes. The real world will move at a digital pace.

Omni-use is the double-edged sword of the coming wave. As mentioned earlier, the same technology that has the power to make medicine can make poison as well.

Autonomy is technology becoming independent. Human intervention and oversight become unnecessary. AI is already a black box that is unable to explain its output. You can take apart and understand a steam engine, but you can’t take apart an LLM and see how it functions, and the paths of logic it takes to give an output. A world like WALL-E where we are unable to take control of our lives because we don’t understand how anything works.

AI and biotech will become general-purpose technologies that will lead to further innovation. Just as the steam engine started the Industrial Revolution which led to production lines, factories, and low-cost goods. Except, as mentioned earlier, it will be at an internet pace. General purpose technologies cause bigger waves, making “rippling amplificatory loops.”

The wave is inevitable because the incentives for it are too strong, despite the almost certain consequences. Competition of geopolitical tech, global research and the pursuit of knowledge, financial gain, and ego drive innovation in the space. It’s common sense to want to develop a way to stop climate change, a no-brainer to want to end poverty, and obvious that you would prefer it was your country responsible for these things. The benefits make the coming wave seem and force us to ignore the potential chaos that might ensue. Suleyman defines this attitude as pessimism aversion.

Knowledge always needs to be sought out for the sake of knowledge. Why did we develop the thermonuclear bomb after we saw what the atomic ones could do? Because we knew it was possible.

Containment

The only semi-successful containment of technology so far has been with nuclear weapons. This is because the development of nuclear warheads doesn’t get much cheaper over time, they are incredibly difficult to create, and the world agreed to stop pursuing the field since it agreed it would destroy humanity if it didn’t. Bombs don’t promise to cure illness or extend human life decades longer as the coming wave does. Containment does not mean a full stop to AI research. If AI development were to completely stop around the world, it would lead to dystopia. Stagnation means the death of society and unfettered progress means complete chaos. Containment is managing AI and biotech development to keep it on the right track, which Suleyman calls the narrow path.

Safety must be a priority. Encouraged, incentivized, and funded so biotech and AI do not create chaos. Companies developing the coming wave must be transparent with the public and with governments. There need to be systems in place to track advancements. Development should also be slowed to allow the world to better prepare for the coming wave. Choke points exist within the supply chain that can be used to slow down progress. Among these bottlenecks are chip manufacturing, cloud computing, and skilled labor. Critics should be on the ground floor of this wave’s development. Complaining on social media makes little difference. While businesses must make a profit, less emphasis needs to be placed on shareholders and more on stakeholders. A commitment to ethics is paramount. Governments can no longer be the slow-to-progress dysfunctional bureaucracies they have always been. Regulation needs to be in place before something happens, not after. Internal and external cooperation between governments is essential; a worldwide alliance to monitor AI and biotech. The culture around the wave needs to change. Failures cannot be hidden. Industry-wide calls need to be made for what is okay to pursue. Developers need to pause and evaluate the potential impacts their actions might have. People must ditch the pessimistic adverse mindsets and understand that change is coming.

Coherence between the above is the best shot we have of staying on the narrow path.

I hope the future isn’t as bleak as Suleyman predicts. I would rather not live through seeing all governments deteriorate and society turn into a series of self-sustaining communes where no meaningful work is done by humans. The steps he thinks are required for society as we know to have a fighting chance seem unlikely. The 6-month pause on AI called for by industry leaders proved ineffective. Elon Musk, who was a big proponent of the call, was developing an LLM, Grok, during the pause that didn’t happen. Asking countries like China to pause AI development in the middle of an arms race seems laughable. If there’s one thing that humanity has always been good at, it’s not creating worldwide unity. I will try my best to remain optimistic and do my part to preserve the humanness left in the world, but I don’t know if it’ll end up making a difference.