The Age of Wonders and Terrors

In 2006, many experts dismissed the notion of AI taking over the world within our lifetimes as pure science fiction. They argued that before such a scenario could unfold, we would first witness warning signs like AI agents breaking out of containment, hacking websites, and solving major mathematical problems like the Clay Millennium Problems.

At that point, it would have been the time to panic. However, over the past two decades, these predictions have come true. The first signs emerged a decade ago with AlphaGo, intensified with the rise of large language models (LLMs), and coding and reasoning agents, culminating in this summer and fall with an unprecedented acceleration of wonders and terrors.

Critics often dismiss these developments by arguing that AI can perform tasks that were previously unimaginable, but they conveniently point to the fact that AI hasn't yet solved the Millennium Problems or escaped its sandbox. They claim to be conservative and skeptical, updating their position based on the latest evidence. However, the author contends that this position remains the one they held in 2006, updated with honesty for the reality of 2026.

The author believes that the Singularity - the point at which AI surpasses human intelligence and rapidly accelerates its own evolution - has already begun, albeit unevenly distributed across different regions and individuals.

The author argues that it would be healthy to acknowledge the reality of these developments, rather than getting caught up in ideological debates. They feel a sense of unease about the future their children may face and struggle to provide guidance to their graduate students about potential career paths. The author draws a parallel between accepting the reality of AI's progress and accepting a religious belief, emphasizing that the former is based on tangible evidence and the latter is a matter of faith.

They acknowledge the uncertainty surrounding the future but assert that AI will undoubtedly play a central role in shaping it.

The author concludes by admitting that they were wrong about the greatest challenge facing civilization, but they updated their perspective once the predicted wonders and terrors started occurring. They urge others to do the same and acknowledge the reality of AI's impact on our lives.