Hacker News Reader

I’m surprised no one has attempted to train crows for use in asymmetrical warfare. At least before the advent of cheap drones.

Pigeons were used in WW1 and WW2 for communication. Paddy the pigeon [0] flew 230 miles across the English channel to relay the success of the D-Day invasion.

[0] https://en.wikipedia.org/wiki/Paddy_(pigeon)

And “almost” as guide systems for bombs! https://en.m.wikipedia.org/wiki/Project_Pigeon

I've known about pigeons being used for communications and even bottle-nose dolphins being used for clearing mines in WW2. But I'm just curious if they've ever attempted to use crows for reconnaissance or early warning systems. Especially for use at something like forward operating bases that are always prone to enemy ambushes. Or maybe even using crows to alert of enemy movements.

I watched a video earlier today on a YT channel I follow called Curious Droid. This episode went into how the U.S military had a hard time determining vietcong troop movements due to the thick jungle foilege. So DARPA developed this concept of electronic fenses, where the airforce would drop these sensor packages into the jungle. The package would have sesmic sensors and microphones to capture movement of enemy forces through the jungle. The problem was that this being the 1960s/70s - the batteries only lasted couple of weeks. Also data storage and tranmission rates weren't advance enough at the time to send that information to a centralize location far from the contested area. So they had to have an aircraft loitering above to collect this data and then fly the collected data to a processing facility in Thailand. By the time the data was classified and analyized, the intelligence collected wasn't really actionable for the commanders in the field.

Here is the video. https://www.youtube.com/watch?v=feDk6oaeVAY

I find it fascinating how scientists still haven't definitively figured out how magnetoreception in birds works. Humans have utilized homing pigeons for thousands of years [1], but it is still a mystery as to how it works. To quote a paper from 2019 [2]:

> Yet in spite of considerable progress in recent years, many details are still unclear, among them details of the radical pair processes and their transformation into a nervous signal, the precise location of the magnetite-based receptors and the centres in the brain where magnetic information is combined with other navigational information for the navigational processes.

1. https://www.atlasobscura.com/articles/ancient-egypt-pigeon-p...

2. https://doi.org/10.1098%2Frsif.2019.0295

I thought I read that they had detected something in the birds' eyes.

Hilariously, during WW2 there was also a research project to build guided missiles using pigeons in the nose cone. "The nose cone of the missile would be split into three compartments, with a lens projecting an image of the intended target onto a screen at the front. A pigeon in each compartment, trained by operant conditioning to recognise the target, would peck at it continually."

https://www.military-history.org/feature/pigeon-guided-missi...

Worms still use them as homing missiles for their Armageddon battles

Hi! What planet are you from?

Didn't they have wireless communication? What benefits did the pigeons have?

Harder to intercept?

You'd want to encrypt your communication in either case.

Triangulation of the sender (or the receiver by way of emissions of the receiving set) was a problem in WWII. There's also the problem of traffic analysis. I would think birds would be immune to both.

This is a bit in a Chris Morris movie called Four Lions.

Barry says we come out blurry

They'd still be better than cheap drones at remote pick-pocketing - if you could train them to spot and snatch coins and bills.

I was wondering about the opposite too. Could you train a crow to watch you throw a marker, say a pebble of specific color, and then fly to the location it landed to perform specific action? Say, if we go for personal mischief instead of warfare:

- red marker = find whatever interesting object there is (e.g. man-made things it can carry, or money), and fetch it;

- blue marker = grab a payload from you and deposit it at location (e.g. a piece of paper, a "stinky bomb")

- orange marker = hover over location for a while, or explore it (useful with e.g. crow-mounted camera; throw marker over a fence or on a roof, get footage of place you can't get to)

- violet marker = make noise, scare away meanest looking people or animals near target (say, to chase away partying college students that decided to occupy the nearby children playground before sundown, or to remind the neighbor blasting his car radio at 5:30 in the morning that there exists small children, or decency, or noise rules) (both totally random hypothetical scenarios, yes); alternatively, to patrol the area for a while and make noise if people show up;

- rose marker = find nearest switch or lever, flick it;

Etc.

Of course the crow would need to be trained to fetch the marker back after executing the action, as to prevent accidental repeat actions, make markers reusable, and ensure there's no evidence left behind :).

I suspect they would be difficult to train, because of their attitude. We see this with many wild animals. Dogs train easily, but coyotes and wolves, who are even smarter, are much more difficult to train.

One of my favorite nature vignettes, is a part of Planet Earth, by the BBC, about Cape Hunting Dogs[0]. It looks like a military assault. They are really clever, and work well, as a team.

But if you could train crows, they would probably be marvelous.

[0] https://m.youtube.com/watch?v=MRS4XrKRFMA

Train is a bit of a shorthand here, though. Basic operant conditioning probably won't cut it. But the idea itself came to me after encountering similar thread about crows on HN many years ago - both comments and referenced articles would contain personal anecdotes about human-crow relationships, and from those I gathered that it's possible to kind of befriend a crow (or a group), and teach them behaviors as a kind of play.

So, I don't expect it's easy to train crows like dogs and pigeons, where one person "pre-trains" them and gives/sells them to someone else. But I think it should be possible to "train" a crow for yourself - bond with it and have it voluntarily indulge you reliably enough that it's pretty much like training.

Badgers are incredibly smart and we seem to not be able to train them at all. There's a pair on YouTube that figure out how to escape their sanctuary

Honey badgers? I visited a place in Africa where a pair were housed in an old tile-sided swimming pool because they could get through anything else. Until someone left a long-handled garden tool in there and they used it to climb out. One of the two was never seen again. The other was caught in a lion enclosure - it had cornered of the lions who was trying to avoid a fight, allegedly.

A classic Honey Badger Meme: https://www.youtube.com/watch?v=4r7wHMg5Yjg

It’s interesting how our view of animal intelligence is massively shaped by the animals we see every day - pets, farm animals, zoo animals. All animals that have had their natural intelligence degraded substantially.

Dogs are wolves. All of them.

The inevitable result: https://cmarshall.com/images/WhatsTheWorst.jpg

Thinking about markers, the falconer at a railway station I frequent paints the pigeons he wants his bird to pursue with a green laser.

> They'd still be better than cheap drones at remote pick-pocketing

for that literal reason, they are illegal to keep as pets in srilanka

ISTR a study where someone tried to train crows to collect litter in return for small food treats. It worked fine until the crows discovered some old pizza and decided it was better than the study's rewards.

Such a human move. Learn something remarkable about nature, next step then is thinking about how to use it for warfare.

Should also establish that the crow on the balcony is the same one that attacked after line of sight was broken during the trip back upstairs.

Not sure it is necessary as the captive crow may communicate it status (croaking "I'm trapped" or "I'm free") with this one, but it would be an insight.

I'm not sure I understand. I'm referring to the assailant crow that allegedly flew around the building and up to the balcony while the original poster was indoors ascending to the eighth floor. I'm not sure how it can be known, given the details in the post, that the assailing crow wasn't a third crow other than the two on the balcony.

About 6 eagles live nearby. I often see them circling overhead hunting wabbits. (Occasionally they'll bag a small aircraft.) One day, I was out for a walk and one of them was gliding at eye level and passed me. I could have touched its wingtip it was so close.

My gawd, what a magnificent bird. I then understood why combat pilots would paint them on the sides of their fighters. It was much bigger than I realized, and the size of those talons - I was not going to get it mad at me!

Raptor talons are really serious weapons - they are the killing mechanism, not the beak. In larger birds comparable to a Rottweiler’s jaw in power. They also have a ratchet-like behaviour, so the bird has to actually want to let you go once you have been grabbed, and struggling usually only makes the bird grip harder

Oh yea, their claws are wild. I got to experience the Tucson Desert Museum demo[0] (it's still going, but this was 16+ years ago)... it's amazing. The falconer had a leather covered arm even though the bird is a close friend.

They are smart AF, but watch out for the beak too if it's injured/scared. A vet told me, if you handle a raptor, protect your hands, they can take a chunk.

Same goes for long billed birds[1].

[0] https://www.desertmuseum.org/visit/rff_index.php

[1] https://youtu.be/MstjYUmdCwo?t=430

yeah, I volunteer at a UK raptor conservancy. For handling most of the birds a falconer's glove is okay but for the bald eagle we have a long leather guard for the whole arm - it's actually a boot that someone cut the foot off. A couple of the handlers describe being seriously gripped by a bird - you just have to keep still and grin and bear it until the bird decides to let go.

I got a scratch on the side of my head once from a black kite that 'clonked' me with its talons as it flew past but this was just messing about, not a serious attack.

One other thing is how I would feed the birds. Some of them (e.g. a hooded vulture) I would pass food to their beak with my bare hand but with other birds this would be a good way to lose a finger.

Also, their Egyptian vulture likes playing with shoe laces (undoing them by pulling) and also tugging on leg hair if you are wearing shorts. The Yellow-billed kites like grabbing people's hats if they are wearing them.

It has been established already that crows remember human faces for years, that they mourn, and that they attack people they associate with the death of their mates (while being quite friendly with other people they know well).

> and neglect any other human walking by this back alley

Not just the back alley. The back alley itself has nothing specifically to do with it. The test would have to be something like the crow attacks this person but not anyone else around the building generally. That doesn't really make a lot of sense from the crows perspective either though... it's concerned with the plight of its mate, so why focus instead on a random person who's a decent distance away unless it can perceive a specific connection, e.g., by recognizing them.

It's moot anyway. I think it's pretty well established crows can recognize human faces.

> Not just the back alley

The idea is that we already are sure they attack there, and therefore limiting the experiment to this back alley may be sufficient: there is no immediate need to extend the perimeter to the whole zone around building, however if the experiment fails it may be considered.

> it's pretty well established

IMHO an experiment gets better with each neglected assumption.

If the crows who live in the area didn't do this for every other day, but only on that ocassion - and stopped when the issue was solved, I don't think we need to establish anything, it's kind of established already.

You’re not wrong, but crows definitely recognize individual people.