Hacker News Reader

Comment by alexose | Hacker News Reader

This "meta game" was my absolute favorite epiphany as a biology undergrad! Genomes evolve in ways that allow beneficial mutations in the future-- And genes aren't so much randomly scattered in chromosomes as much as they are grouped in ways that maximize useful insertions/deletions/frameshifts.

Evolution has its hands on so many different biological processes, including itself. Nature is astounding.

View on HN

Sadly this doesn't work. There is close to zero selection pressure for genes that improve evolution. Because of sexual reproduction, evolution operates on the level of genes. And genes are randomly mixed up every generation. A gene that increases the probability of beneficial mutations doesn't help itself spread. See my other comment on this thread for more specific examples.

This seems like the individual-selection vs group-selection debate. Some genetic behavior, such as altruism to give a crude example, hurt the individual but help the group as a whole. By your argument, all such genes will be weeded out, and the only genes remaining will be those that strictly help the individual. However, the entire field of group-selection is predicated on the idea that such altruistic genes can indeed flourish, through the process of group-selection.

Admittedly this is a contentious topic and there are many people who dispute that natural selection can ever occur at the group level. However, if you do buy the premise behind group-selection, a similar argument can be made for the success of genes that improve evolution.

It's also worth considering that the "marketplace of nature" isn't as cut-throat as you're presenting it to be. There's enough inefficiencies in the ecosystem, to allow for imperfect organisms to survive. Imperfect organisms that have sufficient "slack" to "invest/divert" into genes that don't produce an immediate payout, but do snowball in the long-term.

https://en.wikipedia.org/wiki/Group_selection

> genes aren't so much randomly scattered in chromosomes as much as they are grouped in ways that maximize useful insertions/deletions/frameshifts.

This isn't true. Almost all mutations are either harmful or neutral, so any such gene grouping would be selected against.

It's called the "evolution of evolvability" in most bio literature.

Yeah, even evolution is evolving. We like to think of ourselves as highly evolved, but honestly, we're still extremely primitive compared to what life could be given another two or three billion years to play out.

Unfortunately life does not have that long [1]. A billion years at best.

[1] https://en.wikipedia.org/wiki/Timeline_of_the_far_future#Fut...

Unless you mean elsewhere than on Earth. That would also put different pressures on it too.

You're presuming that life doesn't evolve to the point where it can escape this planet and move to other ones.

Even if we have a horrible nuclear war that kills all humans it will only take 100-200 million years for another highly intelligent species to come back and take over again. We've got around five kicks at the can before it's too late.

Or, perhaps, in 1 million years, humans will evolve to be significantly less-intelligent aquatic animals.

https://en.wikipedia.org/wiki/Gal%C3%A1pagos_%28novel%29

> We like to think of ourselves as highly evolved

Well, that's a mistake right there: All life on this planet has been evolving for the same amount of time[0]. Every organism, every species extant today has already beaten incredible odds to avoid extinction (arguably the only true long term measure of success). None of us are any more or less "evolved" than any other.

[0] Of course, if instead of "time" we consider "generations" to be the correct measure of evolution, than all us multicellular eukaryotes are lagging the prokaryotes so much that we need to describe the number of orders of magnitude we're behind using powers of ten.

Except that we optimize on the timescales that are many orders of magnitude shorter than those of evolution. There was a threshold of intelligence in the past, the moment it was crossed, evolution[0] has lost the game.

That is, unless we decide to forfeit by nuking ourselves out of existence.

[0] - Here meaning non-human-directed part of evolution; technically, all humans and everything they do are part of nature too.

Reply on news.ycombinator.com