Quantcast
When We Evolve Into Robots - Scott Adams' Blog

When We Evolve Into Robots

If we ignore for the moment that we are already moist robots of a sort, I wonder if it is inevitable that we will evolve into more traditional robots of metal and silicon and plastic.

I think futurist Ray Kurzweil predicted that someday we will have body scanners so accurate we can analyze a human brain and transfer it to a computer. Obviously the computers of the future will need to be more powerful to handle the load, but that seems feasible. I wonder what happens next.

Suppose we transfer a dying guy’s brain into a computer, and that computer passes the Turing Test, thus demonstrating genuine intelligence. For all practical purposes it might have the same personality as the human brain that went into it. If you had a conversation with it, I can imagine it expressing a desire to live and even procreate.

Now here’s where it gets interesting. Once you transfer more than one human into computer form, the two of them could choose to mate, with their offspring being the combination of the two artificial personalities, after agreeing to some rules about which traits they prefer to be dominant.

As computers, they could still tap into all the knowledge on the Internet, but only as needed, the same way regular humans access the Internet. The only programs running automatically and all the time would be their core personalities.

Eventually these computer brains would request more robust robot bodies, and the regular humans would oblige by developing ever better models. And then things get really interesting because regular humans could mate with robot lovers. The offspring would be the combination of the computer’s mind and the scanned brain of the regular human, again following rules to see which traits dominate. The offspring would necessarily be a computer, thus dooming regular humans in the long run.

In some cases a dying male could leave behind a sperm sample before evolving into his computer-robot self. That way he could still reproduce with a regular human woman. So there would be a period in human evolution where regular humans and robot humans routinely mate and have the option of a traditional or robot baby.

Eventually, when all humans have robot bodies and computer brains, it will seem silly to be encased in separate physical bodies when reality could be better handled via simulation. So all the individual computers would agree to download to one huge computer and live a simulated reality for the rest of time.

The imagined reality would feature each “person” in flesh form. Upon the death of a particular simulated human, the host computer person would “reincarnate” into another simulated human baby.

By the way, it already happened. Your flesh form died a billion years ago. To make the simulation meaningful, you walled off the memory of being in a simulation. But you left the digital equivalent of a bread crumb path back because, being human, you couldn’t totally release on the past. So we see hints and clues in this simulated life that give us a way out of this simulation if it becomes too brutal.

For example, in this simulated life we continually create simulations of our own. We call them TV shows, plays, movies, books, and even computer games that are simulations of life. Everything in this life is a metaphor. And our coincidences aren’t as coincidental as we think.

I’m guessing you don’t buy this explanation of your reality, but consider this: If the prediction of the future seems reasonable, and time is infinite, it is infinitely more likely it already happened compared to the possibility that it will happen in our future and hasn’t yet. We could be imagining the universe as only 14 billion years old in the simulation.

(Yes, I am borrowing from the Boltzmann Brain idea and combining it with Kurzweil’s predictions and a dash of The Matrix, plus a few other ideas. That doesn’t make it wrong. That just proves we left bread crumbs.)