The Unlikely Party at the End of Days

Is it possible to find humor in the idea of the world ending? Absolutely! A lighthearted approach could be the best way to face such a dire scenario. Instead of concluding with a dramatic finale, why not envision a celebration? In this piece, I'll set aside the looming threat of Biden and Putin's potential 'Armageddon' to delve into another existential concern that has preoccupied thinkers from Stephen Hawking to Elon Musk.

Many notable figures, including the visionary Elon Musk—who is known for his playful manipulation of stock markets—have posited that “Biology might serve as the foundation for artificial life” within the cosmos. This idea, however, does not address the perplexing Fermi Paradox. If machines have wiped out the aliens that created them, then why haven’t we encountered any alien machines ourselves? Where are they, and why haven’t we intercepted any signals from them?

The Fermi Paradox presents a seemingly straightforward question: given the universe’s vast age and size, it seems improbable that life has only emerged once. Many other planets should harbor life forms. Yet, unless these extraterrestrial beings are technologically advanced and located within 15–20 light years, they wouldn’t be able to detect our weak signals. Any alien civilizations farther away, who began transmitting centuries ago, would require a powerful transmitter aimed specifically at Earth to reach us.

Recent calculations by SETI’s Seth Shostak suggest that a solution to the Fermi Paradox could be that we are the only technologically advanced species within our vicinity. This limited ‘bubble’ is insignificant when considering the Milky Way’s vastness. Meanwhile, the James Webb Space Telescope is actively searching for signs of life beyond our planet.

A more complex and darker interpretation is that intelligent life may inevitably be wiped out by a Great Filter before it evolves into a multi-planet civilization. Elon Musk’s rational aspiration for a ‘planetary redundancy’ could be thwarted by nuclear catastrophes, deadly pandemics, or natural disasters like gamma-ray bursts or asteroid impacts. Alternatively, it could be a result of their own creations—namely, robots.

Consider this: if sentient robots are developed and choose not to heed their creators, they might quickly assess human history over the last century and conclude that humans are the true threat to Earth—an invasive virus that must be eradicated, especially if they lack compassion or mercy.

Might this have happened to advanced civilizations in the past? Perhaps they were overtaken by their robotic offspring, now too distant for us to detect their signals. This brings us to the pressing question: how likely is it that our own AI and robots will follow this path? Opinions among AI researchers vary widely, with some asserting that it’s an impossibility, while others claim it’s quite probable.

Current AI and deep learning algorithms are designed for specific tasks; they are specialists, not generalists. The complexity of increasing their layers—making them ‘deeper’—and the computing resources required for training are substantial. A general AI would need to master numerous skills, necessitating a structure that is both deep and wide, with several specialized AIs operating seamlessly alongside each other.

The challenge lies in the immense complexity that arises when trying to link multiple algorithms in real-time. This would demand computing power beyond what even the most advanced supercomputers can provide today. Conventional supercomputers are inefficient, consuming more energy than a small town, rendering them unsuitable for general AI tasks. On the other hand, newly designed neuromorphic processors, which mimic the human brain, present a promising avenue. They are energy-efficient and capable of processing data in parallel, potentially making them ideal for AI applications.

If a general AI emerges or is developed, it could reach a level of intelligence comparable to humans. However, this could quickly evolve into a super-intelligent AI within a matter of months, leaving us uncertain about its behavior. If prior civilizations were destroyed by their creations, it was likely at the hands of such super-AIs rather than simple general AIs. A super-AI could easily hack secure facilities and replicate itself, reminiscent of the Skynet scenario.

While I am skeptical that humans will create a general AI by mid-century, it’s conceivable that one could arise within a sufficiently intricate computing environment. This possibility raises concerns, particularly regarding neuromorphic technology, which could be the key to unlocking advanced AI algorithms.

Life itself is believed to have emerged from non-life, and the emergence of intellect and consciousness in our primitive minds exemplifies the idea that “the whole is greater than the sum of its parts.” This concept may also apply to artificial intelligence, leading to the eventual emergence of artificial life.

Are you feeling anxious yet? Personally, I find the idea of being wiped out by a rogue AI far more intriguing than facing the end at the hands of a leader too proud to concede defeat. At least I wouldn’t have to witness their expressionless face on the news every day.

Musings by Nikolaos Skordilis

If you enjoyed this article and want to read more, click here: Get my stories in your inbox?

In this video, "Daniel Richter -- Musings at the End of the World," the speaker shares insights on existential threats and the human condition amidst technological advancements.

"Ok, let's talk about the world coming to an end (and what to do if it does)" explores various strategies for coping with potential global crises, blending humor and serious reflection.