Only SOME Cockatoos can solve this puzzle. Why? 

It’s cute that they’re slowly getting more comfortable with you since they stick around when you bring out or adjust the puzzles

5 years from now: *Making cockatoos figure out quantum gravity for treats.*

Like others said: I think these birds *can* do multi-stage puzzles, but like you said: each stage needs to be something they learned before. Securing the bar with sticks or dowels, or maybe even rope. Or have them learn with a new puzzle that unscrewing something can lead to a reward (and use a different thread so it only takes like 2 complete turns). THEN integrate that into a multi-stage puzzle. Love these bird videos btw!

One of my favorite things about your cockatoo videos is how they demonstrate that intelligence and problem solving isn't just about finding the one correct solution, but using what you have to do what's best for you. for cockatoos, that includes just tearing the puzzle apart, because they've got excellent beaks for doing so. It'd be silly not to tear an obstacle apart if that's easier than finding a clever solution. Its wonderful 😊

Parrot intelligence is on the level of human toddlers, so you could look at children’s toys for inspiration.

I am amazed at how close they let you get. Must be the years of feed them and giving them puzzles to solve.

Another awesome video Angus. Like you said, these birds do your puzzles by choice, which makes it so much more profound and watchable. Keep going with it mate.

Could you try using dowels to hold the knob lock? Maybe put the dowel puzzle next to this one to establish a sequence

Edweena doesn't have object permenance problems, it's just that she learns crazy fast. Those few seconds with the empty puzzle are enough to convince her out didnt work

Omg this made me realise that I have discovered your channel TWICE???? The first time was a long time ago when I was looking at smart birds, then just over a month ago I found you for 3D printing, now this vid makes me realise it's the same damn guy???? Algorithmically you're definitely doing something right!

I'm not a professional animal behaviorist by any means, but I have a couple of thoughts/potential explanations for some of the things in the video: 1) When Edweena tries really hard to pry it open even though she should know that it doesn't work, that could be an "Extinction burst". Sometimes an animal has a behavior that's well-rehearsed because it tends to work often; if that behavior doesn't work in a specific situation, the behavior starts going "extinct", but there are "bursts" where the behavior comes back, often even more intensely than before. It's like the animal saying "are we definitely sure this doesn't work anymore? let's really try it to make sure." 2) As far as your question/musing (if you will) toward the end about why some of the birds are so into the puzzles - there's a phenomenon called "Contrafreeloading" where animals do prefer to work for the same reward instead of getting it for free. My personal mental explanation is that maybe it's evolutionarily more beneficial to get the same reward while *also* learning and/or acquiring skills than to just get the reward. So animals evolve to find these kinds of problems intrinsically "interesting". I do think it should be possible to teach a bird to solve a multi-stage puzzle like this. It's just a matter of splitting the stages more, so that they have a chance to see when their actions are actually resulting in "progress" in some way. Like you said, having dowels that pull out instead of twist-off things could be a good start. Especially for the birds who already know that yanking dowels often means progress. It's really cool how they generalize stuff they've learned to new puzzles! Then once they know that taking the things on the sides off is progress, they may be able to figure out the twist-off version. Or maybe start even further back - have the bar there, preventing rotation, but not attached with anything. Here's an interesting experiment that I've never heard of anybody trying: what if you rig it up so that a specific sound is played whenever they are making progress? Can they learn to use the sound as feedback that they are on the right track? Maybe practice first on puzzles they are already familiar with, so they don't have to be learning several things at once, and then increase complexity and see if they are using the sound cues. I love your bird training videos! I really only started watching your channel because of these, then moved on to the neat engineering thingies, and only just recently actually got into 3D printing and started watching the "main" videos.

These videos are incredible. More please. My first thought is building on what you have done in the past and replace the side screw-in knobs with wooden dowels they can pull out. I would love to see if the wooden dowels make them remember. Thanks for what you do!

Popeye might have also learned that the top is a fool's game after watching Chad fail to get into it, so he went for the next most interesting thing. He got to see that nothing interesting happens at the top, so learned pretty quickly that it's not worth the attention.

WOW what an incredibly cool video and a gorgeous pair of birds! I’m really curious if the birds can retain context on multistage puzzles if they can’t see progress. So for example, w.r.t. the 2 pegs puzzle… Puzzle 1 - clear front. Pull out single peg to dispense seeds Puzzle 2 - clear front. Pull out two pegs to dispense seeds Puzzle 3 - clear front from seed hopper to first peg. Opaque below first peg. Pull out 3 pegs to dispense seeds. Thus, the birds must understand that the seeds are still making progress even without seeing them.

I don't know if somebody had already suggested the idea, but what if you introduce some sort of trigger linkage mechanism? You can use something like a lever connected by a Bowden cable (like those found on a bicycle). It would be interesting to see what these birds can identify as a puzzle and how far away you can place a trigger before they no longer see it as a part of a system

One thing you could try is having the same puzzle setup with slightly harder elements set up right next to each other. Maybe the first puzzle is the turning wheel with the second puzzle having two dowels that you take out to release the turning wheel.

Pretty sure they could figure it out with some small adjustments. Try replacing the side cogs/wheels with sticks that they need to pull out. Pulling things is their first instinct, and the sticks might remind Popeye of earlier puzzles, especially if they're visually identical to the Dow puzzle. Also, I wouldn't be surprised if Popeye was able to figure out the cog by observing Chad or Edweena, like you said. So if you wanna stick with side cogs/wheels, try unlocking the barrier while Popeye is watching you. Then, roll the main cog/wheel to give him seeds, and put the barrier and its cogs back on to see if he'll figure it out.

I do think that they didn't associate "Turning" the knob with solving the puzzle, but rather shaking it, so maybe you should make a knob that only twists in one way so they can associate the puzzle with the reward

maybe try making the central turn nob and the other turn nobs the same color and without holes. that could give them a hint as game logic. good game design has cue consistency

You could also explain the process of you (Makers Muse) creating increasingly complex challenges for the birds to solve as being "trained behavior."

Great to see an update on your cockatoo puzzles. And as with humans, bullies are usually not the sharpest tools in the shed

I wonder if you could do a more basic screw puzzle where a screw holds down a lid. If they can associate turning the wheel for seeds with turning the screw with seeds, they may be more likely to engage with the screws on the more complex puzzle. What a fascinating series of experiments!

You should try simple multilayer puzzles, like pull out the screws to release the barrier, then upgrade it so the pull out becomes twist. Instead of learning from single layer to multi, it is multi to the same but harder multi.

Try the shape puzzle! See if they can distinguish between shape by keeping a reward behind on certain shape or colour?

Have you tried putting the screws to unscrew in front? That way it mimics the seed screw, and they might try those out to see if they do something. On the side they might not make the connection

I would love to see a puzzle based on having to do another step first, put a basked, remove the cover and then turn the leaver / do action. So instead of "do everything, see what works" see if its possible that cockatoos make a visual connection between action and reaction

The magpies, jackdaws and two crows that show up in my garden every morning for breakfast not only love monkey nuts (roasted peanuts in shells), but they also love dog biscuits. Not on the same level as the sulphur-crested cockatoos, but it amazes me every day watching one of the crows juggling with the dog biscuits and working out the best way to stack three of them on top of each other before flying off with the bounty. I also remember (when I was in Australia), cockatoos congregating around the road houses on the road between Darwin and Jabiru and watching them play tricks on each other - so, just for the fun of it, one bird would sneak up behind another one, grab a leg in its beak and flip the victim over, as if they were wrestling. Imagine a world in which there were only humans. What a horrible thought.

Making the side knobs the exact same color and shape and making them come off with only a 1/4 or 1/8 turn so that they can use the same motion they use on the front knob. Another idea might be to make them turn a series of knobs in sequence.

It might be a big ask but I would like to know if birds can teach each other or even corporate.

put the screw knobs in the front directly in view. so they have visual of it and object perminance isn't a problem

I look forward to next year, where the green bins have knobs to be removed from the sides before we can open them.

a long track along which birds have to push a marble could be fun

Maybe try adding side dowels or printing the side screws purple, like the front one. They might be able to work out that they need to turn the side screws if they look the same as the front one. That is, if they recognise colour?

The birds return! I love this series

Try redesigning the side knobs to look more like the knob that turns the wheel. Plug the hole where the screw was inserted and maybe even add an orange circle behind it. Another thing that might help is to have the locking bar completely cover the wheel that dispenses the food. That could help to discourage fixation. I also wonder if they could figure it out if they saw you assemble the locking mechanism. They're visual learners, but can they figure out how to logically reverse the series of steps they see you perform?

at an old apartment me and my dad used to feed cockatoos and eventually every day there would bel ike 20 coming to visit and they would destroy everything there unless we gave them almonds, some let me pet them if i gave them food and one really old one let me pet him without giving him food i wonder if the new renters liked or hated them lmao

I think the reason one of the birds forgot to use the knob when he was behind it might be related to him not directly understanding that the hopper is the source of the food when rotating the knob, so when he sees the hopper he thinks "I have to break the container" while when he sees the knob he thinks "I have to turn the knob" and the two are not related so it thinks he could do both and starting from one or the other is indifferent

Maybe change the blues for dowels before giving them the blues. Give them a connection that those need to be removed, and perhaps rather than screws is just a simple twist, and it's out.

Thanks!

I would obstruct the knob with a bar like you did in the video, but in a way they cant see it, only a part of the wheel inside and food. As to the side screws- those are tough, could you try with wooden locking pins instead? Those can be pulled out quite easly.

So sweet to see this!

Maybe a sort of elevator design, like the bird has to push a pole all the way up before the platform locks into place at the top and they get the seeds. It's a simple action but its also one that requires continued focus so it's harder for them to solve it by accident.

I find it so interesting that there is a clear divide of Cockatoos, in that passive cockatoos learn skills, while more aggressive ones do not. Makes me wonder on the division, are there more aggressive cockatoos than passive? Are there any outliers, like aggressive and smart wild cockatoos? and if so, are there enough to have their own category or would they be an anomaly?.

Maybe a different puzzle with just the bar (in front of an dispenser opeing that doesn't require turning) that needs to be unscrewed. And after some back and forth between the screwed bar and the twisting puzzles (after they learned and remember these) combine them again. Basically instead of sequential learning, see if they can combine two already known solutions. Just change as little as possible when combining two puzzles.

I'm not familiar with bird psychology, but a multi-step solution might be above their abilities, especially one that is four steps detached from the single step version. If gradually introduced they probably can, but that's just trained behaviour I reckon. Four steps is pretty involved for most animals: 1) unscrew either the left or right screw 2) unscrew the other screw 3) remove the bar preventing the gear from turning 4) turn the gear, get the reward The problem lies in the ability of abstract thought, to calculate ahead. You need to be able to piece together that these three seemingly unrelated actions correlate to the desired action such that the reward can be dispensed. This is a fairly high level of spatiotemporal reasoning as well as abstract reasoning. Realistically this also requires understanding that screws can only be unscrewed one way, I could see how an animal would alternate between tightening and untightening it, or tightening it until it won't go further, or untightening it a few turns only to not see an obvious change and giving up, so an alternative mechanism like a latch may be superior? Hell, a human child might struggle with this mechanism, I don't know at what age it'd become trivial but I suspect around 4? I'm not particularly familiar with child psychology either.

I wonder if Edweena went for the lid on her second attempt because she saw the seeds go in via that route and figured they might come out the same way. Also love that Chad is a dumb jock 😂 I wonder if Popeye noticed Chad bashing his beak into the poly fruitlessly and realized that he needed a different strategy. If so, very impressive observation!

i think you should try a heavy safe like mechanism like with the spinning lock mabey more simplified i would love to see them try

Maybe you could do the trained behavior thing with making several puzzles and then combining them later

Fascinating, Angus. Thank you for sharing.

try things like associating shapes / colours with actions and introduce new challenges in the least "complex" way , then ramping up the difficulty as the behaviour becomes learned. I think the difficulty jump of the turning action with the lock probably stopped any progress. in multi-step problem solving tests try hiding the next action behind the previous, in this case covering the centre triangle with the black piece. if you placed the screws on the front and made it so twisting in either direction loosened it (think of a flag shaped key with a ramp on either side to push it over the "lock position") i think you may have a better chance of them solving it. bonus points if you add a method of dispensing seed to the actions as well

Be interesting to see if you created a cooperative puzzle

I wonder if you started using a systematic color system, red needs to be turned, blue needs to be pulled, etc. if they could associate the color to the task. Therefore you could put more complex puzzles together as there would be an understanding.

Keeping the colours consistent is probably a good idea too, and maybe use more distinct colours for bird's eyes

I would try setting up a puzzle that reinforces that cylinders get pulled, triangles turn, and squares are levers. Then see if they can handle more complicated versions

Passive memory for the win! 😂🤣

I really enjoy these videos, so glad you had time to make another

_"Breaking news: troop of cockatoos have managed to steal the crown jewels. From our reporter in London. [...] It seems they were meticulously trained by an Australian RU-vidr to solve complex puzzles..."_

It's cool to see that edweena and popeye have some problem solving skills that the other birds don't have

That's so adorable 🥰 I remember watching your first video on mazes for birbs, and I am glad seeing you to continue this trend still. About a maze idea. I was fascinated by the way how Popeye waited patiently watching the rogue bully birbs trying to brute force the device and fail miserably, and later uses his knowledge and previous experience to entirely skipp the force attempts and go straight to the lock-picking. And basically speedrun it getting trets in seconds. It got me thinking, that it will be curious to make the puzzle that punishes brute force, and rewards patience and intelligence. Several ideas for this. 1) put either force (weight) or seismic (accelerometer) sensor on a device, to spot if the suspect (birb) is trying to forcefully open the device. If force is detected, scare the offender away by noise, vibration or spinning. The treats will be given only for low force actions like pressing a button or pulling the lever. 2) Same principle of detecting force, but instead of scaring, demonstratively block the device for a certain time. Like rise up the panels that will fully cover the device, or opposite - make the device itself hide by sliding under the surface level. So that the goodies can be only own by doing gentle precise actions without reaching the force limit. 3) Make the patience the solution itself. Put a weight sensor or some other way to detect the presence of the birb. Then start the timer, and if during that time no barbarian behaviour was detected - give the reward. You can even visualise the timer in some simplistic ways, like an hourglass, marbles or some chain reaction, and make it very obvious that force interrupts it's work. It's very curious if parrots can find out that sometimes do nothing and wait is the solution. Also would be curious to leave this puzzle on it's own with plenty of seeds in it, so that it will be fool and bully-proof, and will only feed birds who are patient, persistent and smart. Maybe even behold the creation of the birb secret cult, where only the chosen ones keep the secret, or maybe even teach others they care about on how to use it P.s. Oh, and also, please make some kind of bowl or compartment for seeds this time and not just throw them out to be all over the place 🙃

Are you working with a university or avian research team? I really think your entertaining research in to cockatoo puzzle solving is of greater value for behavioral research.

Popeye: "Don't work out the handle, Chad. Don't work out the handle, Chad. Don't work out the handle, Chad!"

I WOULDN'T EVEN LET THE BULLY HAVE A CHANCE FOR A TREAT

Some years ago I was driving in the country when I stopped at a picnic spot by a little lake. On the lake there were many ducks( about 30) . As soon as we started to get out food, the ducks came as one and encircled us obviously wanting us to throw tasty morsels. I don’t believe in feeding wild animals as it stops them from natural hunting etc. And human food is not good for them. But As I had a couple of apples I thought that this one time (if only to get them to leave for a while, I would give them some. So I bit off a couple of chunks and threw them a good distance away. Again, as one, they ran or flew to the pieces and a fight over them played out for a couple of minutes. That was when I noticed that one duck had not gone. He stood just a bit away and was holding one leg up and very limply towards his body. I realised that with only one leg he never stood a chance of getting anything thrown. It was so sad. Just then the rest of the flock returned to harass us. That is when I formed a plan, I bit off two pieces but I only threw one. They all left as before except the little lame duck. It was then I threw him the second piece of apple and watched with joy as he hopped over to it, and had it down in a gulp before the others returned. I repeated this about 4 time except after that I only pretended to throw a piece to the flock, but they were so conditioned they took off anyway. While they were busy searching I fed the lame duck real pieces. Eventually we had to go and packed up. As we walked to the car the flock started to make their way back to the water instinctively knowing no more food was coming. As we got into the car I turned to check the lame duck was okay…just in time to see him put his lame foot down, walk very briskly to the waters edge and swim off. Now, this duck may have had a sore foot at some stage that caused him to have pity shown by humans and after the foot healed he just used it as a ploy because it worked…a learned behaviour. But I get a strong feeling that when he joined his mates he said, *“see, I told you it would work, they’re so gullible”* …and they all had a good quacking laugh!

makes me want to have cockatoo in my country

Aliens with a quantum physics puzzle: Only SOME humans can solve this. Why?

I used to live next to my boss and his daughter liked a cockatoo.

I would like see a puzzle with some sort of weighted door release...They would need to tug a weight for the door show a hidden prize? Maybe they can smell the seeds and not just have to see them?

the thing is now they know when they turn something, they get to food. but the dispenser works even when rotated only a bit and when turned in both directions. if you would give them someting like a ratchet mechanism they would figure out that sometimes it should be rotated in only one direction. then they would try it at other turnable things too - thats my guess here ^^

I think something with water levels like some of those raven videos I've seen would be pretty interesting

Thorndike was reincarnated and now he’s messing around with cockatoos 😭

Awesome. Australian birds are smart.

Seeds aren't great for parrots to eat since they don't offer much nutritionally and contain high amounts of fat, I recommend pellets instead for something relatively non-perishable

Love all your videos.

Not quite raven and crow levels of intelligence but impressive none the less.

Awesome video! Pavlov would be proud.

I can't get over how wild cockaoos have all but "domesticated" themselves because they realised proximity to humans = easy food 😂 Even pigeons are wary of people despite being comfortable in cities.

What would be funny is ultimately ending up with a puzzle that popeye can open regularly but none of other birds can figure out, so he and only he gets the food. Maybe keep using same puzzle but try and get popeye to figurre out the threads on side or change it to pull off instead of threaded and once he figures it out after some guidance will one day be able to solve it with no issue, while other birds cannot.

this dude is teaching the birts to enter in a house and eat from the refrigerator

Perhaps the cockatoos are looking at the orange circle too? Adding a feature resembling that to the screws might make them realise that those can be turned as well. After which, if given enough time, they will probably figure it out eventually if they don't give up.

How about repeating sequences? Or rhythmic manipulations e.g. adding momentum to pendulum swings? These should test bird cognition of dynamic systems. May be interesting to watch.

i might try these on the local cokatoos

Is there any way to make them cooperate? One of them does something that gives the other one seeds and maybe they'll learn that giving the other one seeds will make the other one help them get seeds. Not sure how this would be done without having a cage for each bird though

I guess in few years you could be using them to solve calculus equasions for treats

yayyyyy another cockatoo video!

Compare the smartness of different birds, not only the yellow, but also the pink and also a kookabara

Some of them got a degree in mathematic.

I bet that Kea‘s can do that too!

Is it too much with several rotations? Like, the triangular sideknobs doesn't provide good enough feedback on a single rotation so they don't know that they need to keep going? The gray triangle gives them seeds with a relatively small turn. I'm thinking perhaps if you can make some kind of latch or lock with the triangular knob that relases after a small turn, and you gradually increase the distance they need to turn, perhaps they will understand the concept of a screw in due time

maybe the result would have been different if you started with the more complex one? or did it in a way that the first stage of the more complex one would have been the original 'get food' stage so that they don't fixate on the later stages of the puzzle.

I know so many humans who find a solution to a problem and life is good. Add an additional level of complexity and the world is apparently ending - like my reaction when we have a software upgrade on the systems at work. I'm not alone. 'Why doesn't this effing work' was heard all around the office for days after the latest upgrade and no-one could get the printer to work.

You could easily market your puzzles for pet market to help domestic birds. They are much safer than the crap we get from China, which is where most of it is made, with inappropriate materials, choking hazards, and ones that will end up with surgery or death. The pet trade doesn't give a crap and they charge TONS of money for them. I try to make my own but I don't have fancy equipment, access to materials or tons of money to order proper materials online. Love what you do and yep, Popeye would figure it out. Maybe for the sides you'd have to take smaller steps, maybe put the same on the sides as the front to teach the same movement is required then the purpose of it changes rather than the action. It's new, different and probably you and a few other factors. Birds are curious. Some captive birds like puzzles, some don't. My tiels don't like toys, just destroy

Great video as always Impressive birds, and a merry Christmas to you and your family:-)

I was going to say you're so lucky to live amongst such a cool species of bird, but really it's your reward for living with all the other things that can kill you.

That's how you grow a brain. Next generation there'll be humanoid Avians !

Instead of using the nuts/bolts to hold the bar in place, why not try dowels? One of them already knows about dowels so it may help learn to pull that out. Eventually, then go back to the nuts/bolts

8:15 I also wonder if he learned from Chad's failure

Parrots will watch to see how YOU do it -- which is why Edweena went back to trying to pry off the top. So, I wonder if he were to let Popeye see him attach and deattach... if he could figure out how to do it himself?

Not sure it has been done, could they learn to match colors to get seeds.

8:19 he was watching the bully bird and saw brute force didn’t work

1) Popeye just spent 10 minutes watching the other bird try brute forcing the puzzle (which Popeye even may recognize as being stronger), 2) you could probably teach them screws if they look identical to the knob in the middle (same type of object, same type of motion)