Hacker News Reader

Comment by adrianN | Hacker News Reader

Von Neumann probes could probably colonize a good chunk of the Milky Way given a few hundred thousand years.

I think that's an overly simplistic assumption, even if we assume they weren't designed by a rational actor who programmed a generation limit into them to keep them from totally converting all matter in the galaxy into probes (an absolute must, obviously).

Life on Earth consists of Von Neumann machines. That's exactly what life is. Yet our life has not converted the entire mass of Earth into itself, and there's still plenty of places where you can't easily find signs of life. And our life has had billions of years. The assumption that Von Neumanns will very cleanly and exponentially increase in number until they are all over the galaxy is too ideal.

It also assumes that all bodies in the galaxy are equally amenable to colonization and conversion into the necessary resources, and that for machines as complex as Von Neumann probes would need to be, there would be no such thing as mutation or evolution over innumerable generations which might halt the process of expansion.

And that most if not all probes in every generation will even find something other than empty space because the universe is very very very big and the likelihood of traveling in any direction and ever hitting anything is very very very small.

Although if you assume the universe is simple enough to describe on the back of a napkin then it probably is straightforward.

That also makes an assumption about the actual ability to make a machine capable of self replication. That in and of itself is a really hard problem as it has to replicate the entire high industrial base needed to create itself which is a LOT of stuff. Just our current technology requires hundreds of processes and chemicals to create critical things like the circuit boards and the components (capacitors, ICs, etc) on it. Then it needs the intelligence and sensors to identify all the constituent sources and address production problems (eg: impurities in source stock) and breakdowns of it's own internal equipment.

Earth is full of machines that are capable of self-reproduction (e.g. bacteria), and some of them are clever enough to build rockets, so I don't think it's such a big stretch to assume that von Neumann probes are possible.

That doesn't get us any closer, unless we stretch the idea of a Von Neumann probe to it's limits and include life seeding. Also we're a long ways away from anything near the scale required even surviving a trip to space much less actually living there.

Yes, this is the essence of the Fermi paradox, and it's often overlooked. 100k years is nothing on cosmological scales.

That in turn makes assumptions about the practicality of interstellar travel etc. But, yes. The counterpoint to space is big is that so are timescales.

This. I always wonder why discussions about the search for alien civilizations don't take the massive timescale more into consideration. Say you're watching a summary of our universe compressed to a hundred years. Each time a civilization pops up, it might just hold on long enough to last a nanosecond. We're just living that in that blip of a time.

Interstellar travel is definitely possible. The only question is how long the trips take. You can build nuclear pulse propulsion vessels fairly easily that can reach a few percent of the speed of light. Those things can cross the Milky Way in less than ten million years.

I wouldn't use the term "definitely." But it looks like "just" a (very hard) engineering problem [i.e. no known fundamental reasons why it's impossible]. And humans are pretty good at solving engineering problems over relatively short periods of time.

ADDED: There are also questions of resources, motivations, and e.g. whether we'd even want to build self-replicating probes.

No, speed is the hard engineering problem. The Voyager probes for example have reached escape velocity from our solar system. That's fairly primitive rocket technology.

Reply on news.ycombinator.com