Distributed Algorithms
December 21, 2016 4:31 PM   Subscribe

Pull Quote II
Most fathers might not be as good at changing diapers as most mothers but, at 3am, the finer points of technique don’t matter.
posted by lazycomputerkids at 4:49 PM on December 21, 2016 [1 favorite]

Technique definitely does matter unless you want pee everywhere!
posted by Joe Chip at 5:34 PM on December 21, 2016 [1 favorite]

Fascinating! But can someone explain this to me:

An ant colony consists of one or more reproductive females, called ‘queens’, who lay the eggs. All the rest of the ants, the ones you see walking around, are sterile female ‘workers’, daughters of the queen and the males with whom she mated.

These males then, where do they come from?
posted by STFUDonnie at 8:23 PM on December 21, 2016

Yeah so the "standard ant" picture (which is not the case in many species!) is that a colony consists of a queen and workers, who are her daughters, for most of the year - no males to be seen. In a certain time of year some of the eggs that would develop into workers develop instead into virgin queens (due to nutrition, worker behavior, provisioning of eggs by the queen), and the queen also lays parthenogenetic eggs that develop into males. Then the newly produced virgin queens and males fly out of the colony on a mating flight and mate (hopefully with individuals from other colonies!). The males then die and the queens land on the ground, shed their wings, and form a colony alone. Eventually they produce a bunch of workers and the cycle repeats. Pretty cool, and complex, huh?
posted by Buckt at 9:42 PM on December 21, 2016 [3 favorites]

Just picturing that scenario with humans. Interesting.
posted by Mei's lost sandal at 10:12 PM on December 21, 2016

ObRef; "Hellstrom's Hive" by Frank Herbert, where a cult of humans more or less successfully emulate socal insects.
posted by happyroach at 11:12 PM on December 21, 2016 [2 favorites]

So, to put it another way:
1. Ants aren't as specialized we thought, they can perform various roles.
2. Humans should too.

ad 1. Well, that's interesting. Thanks! I learned something new.
ad 2. Show of hands, who expects to work one job, play one role in society and family, all your life.
posted by hat_eater at 4:03 AM on December 22, 2016

I've been reading How the Immune System Works, and I was surprised to learn that it's decentralized/distributed in a similar way. After they're created, immune system cells aren't following central instructions. They're pretty much independent operators who communicate and cooperate with each other, like ants in an ant colony. They communicate a lot, in complex ways, again like ants. But their communication is mostly local; it's communication with the cells in their vicinity who can smell the chemicals they're giving off, or with cells who are even closer, in contact with them and "rubbing antennae".

A problem with distributed algorithms shows up in septic shock. In some generalized infections, immune system cells all over the body signal locally that the local blood vessels should get leakier. If this happens in one place, it's a good thing; it allows more fighters to join the fight in the tissues. But if it happens all over, there's a dangerous drop in blood pressure and the patient often dies.

The spread of the 2003 blackout was another example of a distributed system failure. Each system was monitoring local signals - how much power is running down this power line right now? - and each system tripped offline in a way that made sense locally.

Which makes me wonder... are there distributed failure modes for ant colonies?
posted by clawsoon at 9:47 AM on December 22, 2016 [3 favorites]

happyroach: ObRef; "Hellstrom's Hive" by Frank Herbert, where a cult of humans more or less successfully emulate socal insects.

With artificial insemination and surrogate motherhood, we have the technology to do it now. One woman would have to have all of her million or so eggs fertilized and implanted in surrogate mothers. It would be difficult politically, but the technology exists.

If she wanted it to become the new standard for human society, she'd be best off having all her eggs fertilized by the same man.

But that makes me wonder: If humans did that, loss of MHC variety would probably doom us to being wiped out by a single disease. Why don't ants get wiped out in the same way?
posted by clawsoon at 9:56 AM on December 22, 2016

Which makes me wonder... are there distributed failure modes for ant colonies?

I present you the ant mill, also called a death spiral.
posted by late afternoon dreaming hotel at 10:34 AM on December 22, 2016 [2 favorites]

That's amazing, late afternoon dreaming hotel. It's like Winnie-the-Pooh nearly catching a Woozle.
posted by clawsoon at 10:40 AM on December 22, 2016

A good article--for my tastes things in Aeon can go either way but this was definitely one of the things that keeps it on my bookmarks.

I will observe that people who studied ants when central control (by government or corporation) was trendy talked in those terms, now factories are so last century so we get the computer network analogy.

I'd also quibble with "division of labor" model implying control and the distributed system model something else. The processes still determines what ants do. And if anything it's a more disconcerting model for humanity, to present all choices so the outcomes are the same and the shape of the world can match what technocrats or marketers imagine is best.
posted by mark k at 9:55 PM on December 22, 2016 [1 favorite]

But that makes me wonder: If humans did that, loss of MHC variety would probably doom us to being wiped out by a single disease. Why don't ants get wiped out in the same way?

Mostly because for every human, there are about one million ants out there. Most ant colonies only have a few dozen to a few hundred ants, though the ones you notice with a visible ant hill or a cleared area around the nest entrance are usually quite a bit larger. And there are roughly 12,000 different kinds of ants, with more being found every day, so when a disease hits and wipes out a colony or even an entire breed of ants, you probably won't notice.

One example is cordyceps. These are a variety of fungus which targets insects (and humans in some zombie games) and has been known to wipe out entire huge ant nests once one ant was successfully turned into a spore-laden disease vector. But because they're species-specific, an neighbor colony whose inhabitants weren't vulnerable to that strain of cordyceps can do just fine, probably even better than before since there is now less competition.

Other insects which follow the same breeding pattern (one fertile female, single mating flight, all colony members related and likely sharing the same vulnerabilities) also follow this pattern: when Varroa mites infect a hive or a farmer sprays the wrong chemical on a crop the bees are polinating, entire hives die. I've hiked in areas where there are dozens of termite mounds, but no termites because they all died out when some disease hit. The same sort of thing happens with ants, but because there are so many of them, and so much diversity, we just don't notice when a few million die off.
posted by Blackanvil at 1:51 PM on December 29, 2016 [1 favorite]

« Older Ullevaalsalléen 5, Oslo   |   "I have a crazy laugh." Newer »

This thread has been archived and is closed to new comments