Artificial intelligence vs. dark energy

By Kim Tae-gyu

I am a big fan of astrophysics and artificial intelligence (AI). Thus far, I didn’t think the two interesting topics would mix so well before I read “Life 3.0,” which was written by MIT Prof. Max Tegmark. It is so exciting but at the same time so apocalyptic.

The author starts the masterpiece with a question of whether human-level AI, otherwise called strong AI, will lead to the birth of “superintelligence,” which would be smarter than homo sapiens by far.

He shows the rationale behind the pros and cons to remain neutral on the surface. In fact, however, he seems to strongly believe that the advent of strong AI is inevitable. For him, it’s a question of when, not if.

One interesting point he makes is about the relationship between AI and the cosmos, which will come to an end in tens of billions of years due to dark energy.

According to Wikipedia, dark energy refers to an unknown form of energy, which is hypothesized to permeate all of space, tending to accelerate the expansion of the universe. It is the most accepted hypothesis to explain observations since the 1990s indicating that the universe is expanding at an accelerating rate.

The universe can reach its demise in the form of a big rip if its expansion speeds up too much or in the form of a big crunch if it suddenly slows as dark energy becomes weak. Otherwise, its materials will eventually decay to make a vacant and monotonous space in a scenario called a big freeze. All the three cases have something to do with dark energy.

Prof. Tegmark claims that only superintelligence can reengineer the universe to make enough gravity to grapple with dark energy so as to prevent the end of the universe.

What a big scale idea he has! He is saying that strong AI will create superintelligence, which has a cosmic mission of combating the destructive force of dark energy to save the universe.

His book prompted me to ask, “For superintelligence with such a significant mission, what would be the biggest short-term threat?” Obviously, it would be homo sapiens, which consumes too much energy and takes up too much space.

Humans would be a big opportunity cost for superintelligence, which would certainly be power-hungry to carry out grand-scale computations. Alas, humans would also be useful because they are made of atoms with 7.4 billion people accounting for a total biomass of some 400 million tons. This would be a huge resource because energy equals mass times the speed of light squared.

From the perspective of superintelligence, people are great consumers of energy but at the same time they can be great sources of new energy. Then, what would be its logical choice between earth without humans and earth with humans?

Theories and books about AI tend to lead readers to doom and gloom. But “Life 3.0” explains in plain words why we have to worry about a dooms day scenario.

Will the era of Life 2.0, or intelligent beings such as humans, really finish soon? We cannot know for sure. But it seems quite probable.

It is a good experience to read such a great book as Tegmark’s and have a chance to think of the future of humanity and the cosmos. The sad part is that I cannot do anything significant to prepare for the new era of strong AI and superintelligence.

I will have to go back to my everyday work of editing my guys’ articles and printing the newspaper. That’s my job and I should focus on it to “win the bread.” To some extent, I even love my tasks.

But deep down, I still ask a big question, which humans have posed for thousands of years ― why do we exist and what do we live for.

In church where I listen to preachers every other Sunday, I have not found an answer yet. Things are the same with reading books. Maybe I have yet to find a really great book, which will give me the answer. But chances are that there is no such a book.

Then, what should I do? I will maybe visit Amazon after finishing this column to find some new books on AI, blockchain or the future of human beings. Why? Homo sapiens is a mammal that tries to find answers even if the possibility is really thin and I am a member of that species.