Hacker News new | past | comments | ask | show | jobs | submit login
How Bees Argue (overcomingbias.com)
198 points by imartin2k on Jan 8, 2020 | hide | past | favorite | 32 comments

The most interesting thing I’ve ever seen was an argument between ants. As a student I worked on a project to use California grey ants for pest control in peach orchards ‘Formica aerata’ (it’s very effective). This involved capturing wild colonies (digging them up) and breeding them in captivity. To move the colony from the box of dug up dirt to the enclosure, I coated the inside of the box with liquid Teflon (future cancer risk?) and slowly flooded the box, the only way out was a bridge to the new enclosure. Over the next ten hours the ants would compete, some moving eggs and larvae to the new enclosure, some moving them back to the queen. Eventually, when the rising flood had converted enough of the workers to the cause of the revolution, they will make an attempt to move the queen, which she and and a coiterie of anti-diluvianists would resist. The bridge to new enclosure would be a solid mass of writhing black (like Venom from Spider-Man). The queen, noticeably larger than the workers would repeat an escape and climb back over the living bridge to the steadily submerging dirt pile again and again, until eventually overwhelmed she was dragged by the masses to the new home, flailing and struggling all the way.

(Imagining "anti-diluvianist" ants cracked me up, especially because of the double pun.)

Does that mean the ants were unable to communicate the upcoming danger to each other and had to rely on first-hand observation? That's surprising given how much they communicate otherwise.

I don’t know, but the way I thought about it is that I imagined each ant was quite simple, and the fight was a distributed mind weighing the sum of the fractions of reality perceived by each individual

I saw a video quite a few years ago where they dump some ants on a platform, with a string leading to another platform with an area that (apparently) seemed like a decent home. (the below is according to my fallible memories)

One ant wandered to the good area and checked it all out. She then marched back to the others, and bodily lifted another and took her to the good area. They both inspected the area and each repeated the trek to pull in another ant. This continued until "enough" ants were in the new area and they shifted into "build a nest" mode.

It felt very oversimplified for TV viewers, but nonetheless was fascinating - a breakdown of just HOW some simple models can lead to more complex emergent behavior. I just tried to find this video without success (on the bright side, there are a ton of OTHER fascinating ant videos out there)

yes, and that’s basically the model of emergent behavior in many (all?) complex systems, such as human societies. the population exhibits properties that only some of the members exhibit, and some members diametrically oppose.

Could be a security mechanism. Moving a queen would be risky. A large number of ants would be necessary to move the queen if she always resisted, which means a very small chance of false positives or a malicious mis-signal.

I suppose it also means that if she gets attacked half way through the move, then there will be lots of ants around to protect her.

> Does that mean the ants were unable to communicate the upcoming danger to each other and had to rely on first-hand observation

Humans can communicate with each other and yet they behave surprisingly similar to those ants and maybe also often need first-hand observation.

Here's an xkcd about exactly that: https://xkcd.com/1028/

Ants have very simple language. Sure they communicate danger, but how to choose among two dangerous options?

And thus was born the byz-ant-ine generals' problem

...I'll show myself out

Scouts attempt to synchronize a march with an obstinent queen who doesn't easily sign-off on the march. The only way they succeed is to do the enormous work required to carry the queen over to the point of synchronization.

This way rogue ants can't destroy the colony without having 51% consensus on where to move the queen.

I'm reminded of a haiku.

"Worker bees can leave.

Even drones can fly away.

The Queen is their slave."

This is from Fight Club; which is notable because the author was trying to convey that having no responsibilities or possessions was the path to true enlightenment.

This is of course a fallacy as "Joe" the protagonist winds up shooting himself and ends up in a mental hospital. We aren't supposed to sympathise with that notion.

This is fascinating. Are there any cases where the queen was encouraging the move? Or do they always opt to stick with the flooding home until the end? I guess what I'm asking is if it's dependent on the specimen or an instinct.

I heard the process of making decision of bee or ant colony resembles the one in mammal brain cells. Thus, it would have been a visualization of dilemma.

Please tell me you recorded video of this.

Check out AntsCanada on YouTube, he films tons of ants behavior that is really fascinating to watch.

No - was pre digital cameras - but others have since

Is there a video of that exact process? (not sure what to search)

Im on the train on half broken phone, so this is not a well researched response


flooding formicarium

Workers move queen ant

Seem to be returning promising videos

link to one of those kind of videos?


Again on train on broken phone, haven’t watched this but it takes you into the YouTube space of ant keepers


* It is meaningful to speak of animals having "opinions" but it's not clear how those match up with human ones.

* Individual opinion is guided by a combination of random chance and direct perception.

* Individual bee opinions have low success rate, but group consensus of bees has an extremely high success rate.

* Opinions acquired via transmission are more weakly held than opinions acquired directly.

* Wrong opinions are held more weakly than right opinions. I find this fascinating because it means there's some information that would not be revealed without the bees challenging each other.

* Opinions are often never changed in an individual, even if they are wrong.

> Wrong opinions are held more weakly than right opinions. I find this fascinating because it means there's some information that would not be revealed without the bees challenging each other.

This seems to be an emergent effect of a system where every individual scout explicitly communicates a degree of confidence, and the rest of the system can trust the degree of confidence expressed by its individuals so that noise in the actual signal (a scout bee mis-evaluating the quality of a potential home) can be smoothed out a larger number of inputs.

That sounds a lot like the one-armed bandit for A/B testing!

That's an interesting comparison to modern human society, where you can't at all trust how confident someone sounds, because everyone has learned to game that system.

Inspired by the behavior of bees and the Seeley's Honeybee Democracy book, I created and gave a talk, titled Extreme Cooperation of Superorganisms - Four Lessons Humans Can Learn from Bees: https://slides.com/petermoskovits/superorganism/

Do we know how bees encode location data in their dances? Do we have a catalog of their dances? I mean does 3 butt wiggles followed by a head bob mean go towards the sun until you smell lavender?

>Waiting another hour or so gives enough time for scouts to return from other sites, and then the entire cluster heads off together to this new site

What happens if a bee is late coming back and everyone is gone?

This happens regularly, the bee cluster doesn't wait for late comers. There are some scouts that keep going there and back but there are also scouts that do that with the "Correct" destination, so they "lead" the stray ones to their new home eventually. In the end most bees end up where they're supposed to.

It's much worse when a beekeeper takes the swarm tho. Speaking from experience, you can see "lost" scouts going to and from the temporary cluster from their destinations for days after you moved the cluster into a new hive. The reason is because there's noone going from cluster to new hive since no scout found that destination. It's a bit heartbreaking.

I collected the "leftover" mini-clusters 4 times in a row coz I felt responsible...

I'm trying to make this DIY beehive https://www.ikkaro.net/diy-beehive/ and I hope to observe his behavior soon.

Borrowing cryptocurrency terms, this seems to be a Proof of Excitement consensus algorithm

Nice to see overcoming bias post. Do people still read LW?

Applications are open for YC Summer 2023

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact