Don’t Worry, We’ve Got This…

The good news, and there is so much good news, is that we have entered the most exciting period in human history. Two transformative technologies are arriving at once. Artificial intelligence is teaching machines to think, and CRISPR is teaching us to edit the genome of any living thing we can get our hands on. Both are progressing rapidly. Both are being deployed at scale. Both are being managed, we are assured, with the seriousness the moment demands. By whom, exactly, remains a charming mystery, but the seriousness itself is not in doubt.

Some people worry. This is understandable. Some people worried about fire, electricity, and the printing press, and history has not been kind to them. The lesson of every major technological transition is that the people warning about it were wrong, except for the ones who were right, who are usually only identified retroactively, in documentaries, after everything they warned about has happened. Until then, they are alarmist. After, they are visionary. The trick, of course, is to be the second kind, but you cannot know which you are until it is too late to matter.

So we proceed. This is what civilizations do. We proceed.

It is sometimes said that powerful technologies are irreversible. This is a soft and emotional claim. A more rigorous definition is that a system becomes irreversible when new causal dependencies form and the prior system reorganizes around them, such that the former state can no longer be recovered even if the new element is removed. By this measure, both AI and CRISPR are quietly crossing the threshold while the public conversation focuses on whether chatbots can write decent sonnets and whether babies should have blue eyes. We are arguing about the wallpaper while the foundation is being replaced.

But this is fine. It is fine because we have a plan. The plan, as best as anyone can determine, is that things will work out. They have worked out before. There is no specific mechanism by which they will work out this time, but the historical track record of things working out is, on the whole, encouraging. The Romans had problems and the Romans are no longer with us, but in their place we got the Renaissance, which had pleasant paintings. By analogy, whatever replaces our current civilization will presumably also have pleasant paintings, assuming anyone is left to paint.

The truly elegant feature of the present moment is the convergence. AI gives us the power to model biological systems with unprecedented depth. CRISPR gives us the power to act on those models. The gap between observation and modification, which once required decades of careful experimentation, is collapsing into a tight feedback loop. Soon we will be able to edit systems we do not understand, in real time, based on predictions we cannot verify, with consequences we will only detect after they propagate. This is widely considered an exciting development.

Critics suggest this might be a problem. Critics, however, do not understand markets. Markets reward speed. The first organization to deploy a recursive optimization loop without adequate safeguards captures enormous value, while the organizations that pause to think about consequences capture market share approaching zero. This is sometimes called a coordination problem, but a more accurate term is “Tuesday.” The structural incentive to move faster than your understanding is not a flaw in the system. It is the system. Pointing this out is impolite.

There is a comforting myth that wisdom will eventually catch up. Wisdom, after all, has always caught up before, in the sense that civilizations did not technically collapse on the exact day people predicted. Wisdom accumulates through experience, reflection, error, generational memory, and long uncomfortable conversations conducted over many decades. The feedback loops we are now building close in a matter of months. The mismatch is sometimes described as a “challenge.” A more honest description is that we are pouring liquid nitrogen into a thermos designed for soup and expressing surprise at the temperature.

Fortunately, the people most directly involved in building these systems have assured us that they take the risks seriously. They have written essays. They have tweeted. Some of them have publicly worried, in interviews, while continuing to do exactly what they are worried about, which is the most sophisticated form of moral concern available under late capitalism. It is a kind of contemplative practice. They reflect in the morning and ship in the afternoon. It would be cynical to suggest this is incoherent. The proper word is “complicated.”

Regulation, we are told, will save us. Regulators are even now drafting frameworks. The frameworks will be considered, debated, weakened, lobbied, exempted, and eventually passed in a form that addresses the previous version of the problem, which by then will have been replaced by a newer and more interesting version. This is not a failure of regulation. It is regulation working exactly as designed, which is to say, several years behind the thing it is regulating, holding a clipboard.

The most reassuring development of all is the public discourse. Across newspapers, podcasts, and panel discussions, we are asking exactly the right questions. Will AI take our jobs? Will designer babies create inequality? Are the machines becoming sentient? These are excellent questions, in the way that asking whether the band on the Titanic was playing the right song is an excellent question. The deeper structural question — whether any institution we currently possess is capable of metabolizing recursive feedback loops at the speed they now operate — is generally considered too abstract for prime time.

None of this is to suggest catastrophe. The technologies will produce extraordinary things. AI may help us understand diseases that have evaded researchers for generations. CRISPR may end forms of suffering that have shaped human life since the beginning. The benefits will be real and they will be enormous, and they will arrive alongside the costs, which will also be real and enormous, and we will tell ourselves a clean story about which were worth which, sometime around 2050, in a documentary.

The irreversibility, when it is acknowledged, is usually softened with the observation that humans are adaptable. We are. We have adapted to fire, agriculture, cities, electricity, antibiotics, nuclear weapons, and the internet. There is, however, a small distinction between adapting to changes that arrive over generations and adapting to changes that arrive faster than the meetings called to discuss them. We will adapt to this too, presumably. The only question is what we will look like when we are done, and whether we will recognize ourselves in the mirror, assuming we still have mirrors, and assuming the mirrors still show the kind of thing we used to mean by “us.”

But none of this is reason for alarm. The dependencies are forming. The loop is closing. The prior world, the one structured around human cognitive bottlenecks and biological patience, is quietly being replaced by one structured around machine prediction and intentional editing. By the time the transition is visible to the people living through it, it will already be complete. This is, technically, how transitions of this kind have always worked. The historical record is consistent on the point.

So relax. Civilization has always landed on its feet. There is no reason to think this time is different. Except possibly the speed. And the convergence. And the recursion. And the fact that the institutions we would normally rely on to absorb the shock were designed for a different century and have not been updated. But other than those few small details, which are admittedly the entire problem, there is no cause for concern.

We have got this.

Probably.