You really should release parts as parametric or at least the source files. I see everything is an STL and STLs are just a pain to work with. Suppose we want to try with mice? Or what about my cat? I do not expect just scaling in my slicer is going to end up with a good result, I'll need to redo everything from scratch. But parametric parts? That gives us a lot faster iteration. That gives you a lot faster iteration too! I highly recommend taking that approach when designing and I find it is worth it more often than not.
Could you add cost estimates to the BOM? These never need to be accurate but I always find it helpful when estimating a project. You're just saving people from the time it takes to click every single link and throw them into a calculator. And informs people very quickly what to innovate on to drive costs down. (Sorry, BOMs without cost estimates are a big pet peeve of mine)
# Questions:
- Do the rats enjoy playing Doom?
- Are there specific games the rats like to play?
I've never thought about what types of videogames other animals would enjoy, but damn if you didn't just open Pandora's Box here. I actually think we could learn a lot about them (and even their specific personalities) from this question. It gives a whole other level of refinement than just knowing what my cat's favorite toys and games are...
And also, thanks for open sourcing this! I'm excited to see what comes of it!
Gonna be honest here, I've worked on this for so long, so many iterations, lots of versions for each 3D part, software and all, at this point I just wanted to publish everything I had and do it fast. And you are totally right, publishing without parametric source files was a mistake, I'll upload everything I have shortly, prices included. Note: mice require smaller setups and that just leads to the redesign of most parts - smaller ball, ball driver, lever with weaker springs... training cats prompts for a larger ball, same issue. VR setups for cats though would be super cool!
On this setup my rats were only habituated, they did not end up playing Doom. Even habituation seamed super slow, they were a year old when I started it. On the previous setup though, when they learnt to run on the ball and how that influences their reward, they got hooked. I believe they enjoy not just the reward, they get a sense of how their actions influence the game and they like that. They would run on the ball so much at some point they wouldn't even bother drinking all the juice and it was just dripping on the setup.
No idea what they would best like to play. It needs to be a first person game though, that's what they are able to understand how to handle, it's more natural to them.
Thank you for taking the time to give feedback! I also hope pet VRs become a thing and people can connect with their pets virtually too!
Is the graphics rendering modified to output actual barrel projection to match the display, or did you just take the original perspective projection and stick it on a distorted screen?
Consider making the screen panoramic around a larger radius rather than just around the head, perhaps on the order of ball diameter. This reduces the visual stereodisparity mismatch and lowers the cognitive load for habituation.
Consider also making the trigger chin- or bite-activated to allow simultaneous shooting and moving.
Oh, I hope you don't take this as me being upset. I'm super happy and totally get the motivation. I a fan of the adage "better to do something half assed than no assed" (not that this is half-assed). Just wanted to make the comment to help drive motivation and let you know there's a demand. Releasing the sources could really help too just so people don't have to work with the mesh.
But on the rat part, that is super interesting! I was suspecting they might not like Doom because shooting a gun might be such a foreign concept to them that it breaks immersion. But it seems like you say they like running around in the simulated environment? (Time for Cheeze-Doom? lol)
Again, super cool and thank for releasing things! This is that crazy stuff I just love to see people exploring.
Thank you! <3 On shooting: exactly that, it is so foreign to them, I doubt they could grasp the concept, but they can understand the loop of: pull lever -> audiovisual feedback of shooting with monster disappearing -> reward. Biting or scratching a surface as a form of attack may work better, but the audiovisual + reward response should help them to understand at what visual signals to pull the lever to make it go boom.
They missed a real opportunity for "Omelette Doom Fromage" there
But yeah, I'd wager too subtle. I'm also questioning now how much rats use smell for navigating their environments. I notice that my cat is a lot more smell oriented than I initially thought and I think it makes a big difference. Hard to tell though.
Humans are /extremely/ visual compared to other animals: this tends to make us underestimate the intelligence of other animals (when we use visual intelligence as a proxy for general intelligence) and miss out on smart uses of other senses entirely.
Rats are well-known for thwarting maze studies using things like fine sensitivity to slope, directional orientation using smell gradients across a room, or detecting the direction of researchers outside the maze based on micro-vibrations.
(Good book on the general topic of measuring animal intelligence: "Are We Smart Enough to Know How Smart Animals Are?" by Frans de Waal)
Funny enough, so does my green cheek conure (small parrot about the side of a fist). Their beaks are made from keratin, like our nails, so it's conductive when touching the screen.
The hiss of the bombs gets him a bit angry though. Parrots hiss and it kind of sounds like that.
As an evolutionary cousin of a rat, the half-second delay after firing my weapon and the response would make me want to eat my whiskers.
This would completely kill any potential reward (and replace it with the opposite, frustration) you're trying to train me with, please fix immediately.
Lol you are correct! At least training them becomes smoother the faster you administer reward. Learning happens at different timescales in the brain, and immediate feedback (about <300 ms) yields the most reliable neural updates.
One of the most important traits of intelligence is understanding and appreciating delayed reward.
I saw a gambler win the jackpot. He was really excited and started gathering up all the chips he'd won. Why he was so excited to win a bunch of plastic chips, I'll never know. What's so great about plastic chips? Why was his brain so excited when all he was doing was gathering plastic chips? ...
A half-second delay doesn't mean your brain can't learn to make the precursor feel good.
I think that’s the same idea behind clicker training for dogs. There’s a delay in giving them the actual treat, but the instantaneous click sound lets them now they did the thing that results in a treat
not pavlov conditioning -- in Skinner's three term contingency, the stimulus context acquires meaning/significance related to the consequences of a response. a neutral or even negative stimulus (context) can become it's own reward through this process. this conditioned stimulus explains most animal and human experience. Humans are especially prone to constructing meaning based on the primal.
Think of the senses: sound becomes talking, music, etc. food become cuisine, obesity, and anorexia. eyes becomes art, movies, etc. desire becomes porn, s@m, etc.
meaning is constructed, socially constructed, or what skinner call "learning." His masterwork, long forgotten, is the "generic nature of stimulus and response." Generic as it open to near total manipulation
i think its just people being out of touch with reality. perhaps engineering minds not thinking enough product. its too commonplace for me to even complain about. someone builds something primarily visual - a robot, a GUI application, etc. and links to their github/etc but they ensure that there are no visuals included.
This is far from primarily visual, but I do understand your point. I could not take videos unfortunately. I mentioned it in other places, but this project took so long, I just wanted to put it out there and get some feedback. I'm glad that people are this receptive to it, and I hope someone would take the project over!
We reached the point of rat habituation but didn’t start training. Our rats (Todd, Kojima, Gabe) aged out before full testing. The setup works, but behavioral validation is pending.
I'm opposed to this project because it involves using animals in medical experiments, which I believe is never ethically justifiable. It goes against basic moral and ethical principles regarding animal treatment. If the project were designed to allow animals to choose whether or not to participate, it would be more acceptable. Some scientists have already explored such approaches. By not giving animals a choice, you're limiting their freedom and potentially exposing them to physical or psychological harm through your simulation. As someone who advocates for animal rights, I'd prefer to see alternative methods that don't involve animals or allow them to participate voluntarily
You have interesting points but I feel you are jumping to conclusions. 'Curiosity not judgment' is a great rule for me, at least, because I rarely know enough to judge. Criticizing without knowing almost certainly makes me wrong, produces nothing, and costs me the opportunity to do something useful. It also wastes time and energy for multiple people.
It also is unfair to the experimenters and alienates them, when they could become allies and improve their methods. It alienates others; it makes you seem defensive and someone who lashes out unfairly - who wants to be involved with that? Even if the researchers agreed, would they want to have this judgmental, attacking person around?
For example, someone could ask: 'Hi - This is quite innovative. How are the animals introduced to the setup, trained, and experimented with? Are they basically required to play? What if they stop? Do they want to stop at the end of the session? Do they seek it out? Are there signs of stress or enjoyment? There is a bunch of innovation in animal research on giving them choices, and as we learn more about animal emotions and intelligence it makes much more sense to consider these things. This experiment seems like a perfect setup to explore some of those things; I'd love to engage with you on it, and/or here are some links to learn about it ...'. The researchers might love to help.
Maybe you know all that. I just hate to see good causes turned into alienation.
> Even if the researchers agreed, would they want to have this judgmental, attacking person around?
1. You calling the person above judgmental and attacking is not as tolerant either.
2. What about things that are morally wrong? Slave owners wouldn’t want to have a judgmental attacking person around either. Does that mean we have to have curious-discussions about slavery?
Your comment finds ways to make it adversarial. What are you trying to build?
> What about things that are morally wrong?
That takes possibly the most certain path toward evil: I think what they do is morally wrong so I can act without morality toward them. It's the rationalization of many awful acts and people and ideologies. Just look at the worst of what religions do.
We should not think we are somehow above or exempt from that error, or so above sin generally that we can preach.
Given that we're not great at communicating with animals, what level of reward would be considered justifiable for participating in the study?
As in they could get a reward for starting the study...
Also, isn't this sort of "voluntary" testing a little unethical in itself? For example, testing an addictive drug on a rat, they don't know the downstream consequences since there is limited communication, but the immediate effects might be incredibly gratifying. It would lead to high "volunteer" rates but still expose them to massive harm.
"when they learnt to run on the ball and how that influences their reward, they got hooked. I believe they enjoy not just the reward, they get a sense of how their actions influence the game and they like that. They would run on the ball so much at some point they wouldn't even bother drinking all the juice and it was just dripping on the setup."
I agree with you, but I would add that even so, it is still far better than 99.99% of scientific animal experiments, in which animals are usually subjected to severe suffering (for example, through the injection of human tumors) and are almost always killed at the end, as their tissues have to be examined.
I struggle to see how it's a acceptance that it's done at someone's home without the animal experiment safe guards and ethics oversight you'd get at a research facility.
Topped off with a plug to get your email address for their new startup.
I rolled my eyes at this at first but after looking at the setup they have I have to agree with you… there’s something viscerally horrifying about hooking something sentient up to a trackball with a VR “headset” with no way out.
It’s something about messing with reality. Obviously I can’t know this since (as far as I know) I am not a rat, but I have to believe it’s profoundly disorienting for their little rat brains to interact with VR. At least a work horse can trust its senses.
>We reached the point of rat habituation but didn’t start training. Our rats (Todd, Kojima, Gabe) aged out before full testing. The setup works, but behavioral validation is pending.
Ah man, what a pity. That VR rig is awesome, but it doesn't really seem to me they are planning to continue these experiments, or do they?
How could you go to this amount of work and not continue??? Probably just waiting for other people to get mad at the lack of progress so that they attempt this themselves, and then he can challenge them to a deathmatch with his highly trainer ringer rats
When rats gaming championships become a thing, I will have the top team for sure!
Seriously though, this project took way too much time. We built 2 versions of the setup, lots of hardware, software building, testing. I hope a behavioral lab or hardware enthusiasts would take it over and scale it. I did learn a lot in the process and would gladly support anyone who would continue it. In fact, I have a setup now laying around in my living room up for the taking :D
They also seemed to basically do it for two weeks and then stop because the rats "aged out" - despite it supposedly working well and achieving concrete results.
I know rats aren't long-lived, but I would be interested to know how they determined the rats 'aged out'.
Could also be a complete failure they spent considerable effort in with zero results, and are hand waving and constructing a way to quit while claiming success.
I could see the rats not really connecting things and just puttering around. It's a pretty involved setup and I poor uptake on the part of the rats would be a steep disappointment.
The year is 2034. Countless attempts at re-producing the sophisticated wetware of the brain have failed. Modeling research has proved unfruitful, with the curse of dimensionality afflicting every attempt at breaking the walls of general intelligence. With only a few million of capital left, and facing bankruptcy, they knew that only one option remained.
I think there is probably an anthropocentric flaw in the rat display.
Human eyes are side-by-side and forward, with a big binocular overlap and a clear vanishing point. A forward-facing curved screen fits that geometry well.
Rats' eyes, by contrast, are lateral. They have a much wider field of view, a tiny binocular zone, and use motion and contrast more than neat perspective lines. A single human-style "cinema screen" isn't laid out for a rat's optics or brain.
Perhaps if the scenes were rendered with a much wider, 250° FOV, it would help the rat understand what it was seeing better.
Or even rendered with two virtual cameras offset and angled apart, then stitch their outputs into one extra-wide view wrapped onto the curved display. That would approximate the rat’s much wider horizontal field of view and reduce the mismatch between where its eyes are actually looking and where the important visual information appears.
There are other differences in perception of color and motion, but fixing the FOV would be an immediate and relatively easy software fix.
A huge shortcoming with this setup is that rats don't just see in front of them, they see all around them. The screen right in front of his nose doesn't really simulate his environment the way it does for a human. Luckily doom has command line options for screens on the left and the right. But I imagine a lot of his awareness comes from what's above him.
Doom aside (very impressive), I love the concept of putting rats on a roller ball instead of in a wheel for exercise. It would be better on their backs.
In the year 2032, the rodrone wars broke out. What started as innocent video game experiments, had taken a dark turn. After the rats had perfected playing the classic Doom games in 2026, they were easily trained on more complex simulations. In 2028 the first rodent controlled surveillance drones were tested. The year after weaponised. And the first real deployments in warfare in 2030. They easily outperformed OpenAIs latest battle systems. Home robots were soon after remote controlled by rats, known as Rodots. Rodent intelligence escalated quickly as lab selections bred only the smartest specimens. It was only a matter of time before the takeover. Now rodots were building steadily more capable drones and bots. Long before humans could foresee the need for a defense against the rats, the rodrone attack on all humankind, was a fact...
Pets.com buys 8M sq feet of datacenter realestate in a deal with Oracle's liquidators, which is reported to include fitting the racks with hamster wheels and feeders. Sets sights on 400B IPO
Here is a thought: Maybe we are all just living in some gigantic experiment, that some incomprehensibly advanced civilization has created, to make us indirectly play DOOM, without us even realizing. Any progress we make is just one step further in their experiment, a reward, every issue like war, climate change, corruption, put in place by them as a challenge, while behind it all, in higher dimension, it is translated all into input for their version of DOOM, while they are watching us perform for them. One of them found it funny or interesting, to give us our own DOOM to play with, and now their scientists are speculating, whether it is always the case, that when you give a test population a DOOM, that they will pass a DOOM onto other beings, recursively.
I believe the practical conclusion is right there and pretty simple: if so, as long as we can't break out of this open world simulation and/or decisively prove we are in one, the best we can do is aim for success in the simulation. Just like we should do in the non-simulated world.
Bad news: no video of them playing on this setup, just on the previous version. We iterated on v2 too long, our pet rats grew old and couldn't be trained. We open-sourced the hardware and software so others can build upon it. You can TLDR the whole thing in this thread: https://x.com/yolorun_capital/status/1996632980903620886?s=2...
...Also, here's my personal X, dm me if you have any questions, or would want to build it for a lab or for yourself: https://x.com/viktor_thoth
I was worried this was going to be one of the petri dish of neurons type things, but the reality of it was awesome. More gaming rigs for rats need to be developed - it looked like Todd was actually enjoying it, and you toned down the potentially scary bits in a good way.
Kudos for the experiment and giving your pet an awesome enriching environment!
<<< just to verify, yes, he is The Rat Coach. Reach him out if you want to build something cool.
(first post we made on the project: https://news.ycombinator.com/item?id=46151150 )
I did not expect this to be as thoughtful as it is. Kudos to the thought put into the setup and training. Building custom hardware for the rat is amazing.. I need something like this for my own setup.
I think it might be _unethical_ to not spread the joy of playing Doom for the first time? Though, I’m not entirely sure there’s been enough research done about the effects of violent video games in rat gamer populations.
We've been putting rats in skinner boxes for a lot longer than we've been subjecting human gamers to them. I'd be more worried about the health effects of all that sugar water.
Giving them slightly acidic water off-rig and normal water while playing is another option - not the best either. I opted for sugar in the end, because they didn't spend much time on the setup; but this should be reconsidered if they on-rig for hours daily.
It's more like locking humans up in the matrix. Note lines like "preventing it from leaving the apparatus" in the build guide. Would an ordinary gamer be restricted from exiting such a contraption?
I was very happy to see this. I’m fairly against live animal testing, but giving rats the joy of playing Doom??? I think I _may_ have to be OK with this.
I haven’t read/watched Clockwork Orange, but I can understand the dystopian image.
It’s just that, when I think about it, I don’t see a difference in kind between training the rat to play a video game and teaching a dog to roll over and play dead. Neither is natural animal behavior, and the animal doesn’t really have a choice in either case.
Very cool... but why not use something with higher res graphics, if you're going to go through the trouble of building the mechanical system. You could go ultra 4K and see if a rat begins to see... reality.
Is this related to the Tank lab experiments at Princeton?
I’ve seen this before with the practical application being that you can keep the animal stationary while it thinks it’s moving, in order to record brain activity.
Wonderfully crazy. Should the rats have multiple weapons at their disposal, which do they prefer? How many demons escape their fate? Are there videos of gaming sessions? So many questions!
yeah good point! we actually have air puffs for that. we extracted from DOOM when a wall is hit in-game. when that happens, the corresponding side puff activates.
So true! A little over fifty percent of the rat population would rather watch fox news and spend more time oiling up those weapons for when "it hits the fan".
Whats the fascination with Doom? I keep seeing running Doom on this and that and now Rats playing Doom. I mean ok, Doom runs on everything and now rats playing doom. So what?
Doom has some star power and is instantly recognizable to pretty much anyone, not just video game fans. Its engine is Free software that's easily adaptable to pretty much any input or output method. In this case Doom might also have made a better candidate than another first person shooter because you can't aim up or down in Doom, so there's no need to account for that in the rats' input.
Doom pioneered the FPS genre in the USA. So it has a cultural significance. I used to go to my friends house to play it. You deal with a whole range of aspects such as grapchics, sound, networking,.. Etc. So it's a fun challange for hackers to try to get it to work in the most strange of ways. In addition it's a simple goal (with a complex path). So it helps with the focus, and most importantly the fun programming challenges involved in making the game work.
one of the first widely known "3d" shooters. Doesn't require much in the way of hardware and there are lots of open source implementations , tools and mods for it. And it's a meme
I absolutely love this!
# Suggestion:
You really should release parts as parametric or at least the source files. I see everything is an STL and STLs are just a pain to work with. Suppose we want to try with mice? Or what about my cat? I do not expect just scaling in my slicer is going to end up with a good result, I'll need to redo everything from scratch. But parametric parts? That gives us a lot faster iteration. That gives you a lot faster iteration too! I highly recommend taking that approach when designing and I find it is worth it more often than not.
Could you add cost estimates to the BOM? These never need to be accurate but I always find it helpful when estimating a project. You're just saving people from the time it takes to click every single link and throw them into a calculator. And informs people very quickly what to innovate on to drive costs down. (Sorry, BOMs without cost estimates are a big pet peeve of mine)
# Questions:
- Do the rats enjoy playing Doom?
- Are there specific games the rats like to play?
I've never thought about what types of videogames other animals would enjoy, but damn if you didn't just open Pandora's Box here. I actually think we could learn a lot about them (and even their specific personalities) from this question. It gives a whole other level of refinement than just knowing what my cat's favorite toys and games are...
And also, thanks for open sourcing this! I'm excited to see what comes of it!
Gonna be honest here, I've worked on this for so long, so many iterations, lots of versions for each 3D part, software and all, at this point I just wanted to publish everything I had and do it fast. And you are totally right, publishing without parametric source files was a mistake, I'll upload everything I have shortly, prices included. Note: mice require smaller setups and that just leads to the redesign of most parts - smaller ball, ball driver, lever with weaker springs... training cats prompts for a larger ball, same issue. VR setups for cats though would be super cool!
On this setup my rats were only habituated, they did not end up playing Doom. Even habituation seamed super slow, they were a year old when I started it. On the previous setup though, when they learnt to run on the ball and how that influences their reward, they got hooked. I believe they enjoy not just the reward, they get a sense of how their actions influence the game and they like that. They would run on the ball so much at some point they wouldn't even bother drinking all the juice and it was just dripping on the setup.
No idea what they would best like to play. It needs to be a first person game though, that's what they are able to understand how to handle, it's more natural to them.
Thank you for taking the time to give feedback! I also hope pet VRs become a thing and people can connect with their pets virtually too!
Is the graphics rendering modified to output actual barrel projection to match the display, or did you just take the original perspective projection and stick it on a distorted screen?
Consider making the screen panoramic around a larger radius rather than just around the head, perhaps on the order of ball diameter. This reduces the visual stereodisparity mismatch and lowers the cognitive load for habituation.
Consider also making the trigger chin- or bite-activated to allow simultaneous shooting and moving.
Oh, I hope you don't take this as me being upset. I'm super happy and totally get the motivation. I a fan of the adage "better to do something half assed than no assed" (not that this is half-assed). Just wanted to make the comment to help drive motivation and let you know there's a demand. Releasing the sources could really help too just so people don't have to work with the mesh.
But on the rat part, that is super interesting! I was suspecting they might not like Doom because shooting a gun might be such a foreign concept to them that it breaks immersion. But it seems like you say they like running around in the simulated environment? (Time for Cheeze-Doom? lol)
Again, super cool and thank for releasing things! This is that crazy stuff I just love to see people exploring.
Thank you! <3 On shooting: exactly that, it is so foreign to them, I doubt they could grasp the concept, but they can understand the loop of: pull lever -> audiovisual feedback of shooting with monster disappearing -> reward. Biting or scratching a surface as a form of attack may work better, but the audiovisual + reward response should help them to understand at what visual signals to pull the lever to make it go boom.
This mod includes collecting cheese (though it's a side activity, and maybe too subtle for the rats): https://youtu.be/qPRvw6kRN-8?si=j9iuTuiHerm0AhQ2
And here's a thing I knew had to exist: a doom mod/level set on a moon made of cheese... https://youtu.be/XxdeUbE9kvw?si=_cpJQKuDy87BN7EP&t=10m20s
They missed a real opportunity for "Omelette Doom Fromage" there
But yeah, I'd wager too subtle. I'm also questioning now how much rats use smell for navigating their environments. I notice that my cat is a lot more smell oriented than I initially thought and I think it makes a big difference. Hard to tell though.
Humans are /extremely/ visual compared to other animals: this tends to make us underestimate the intelligence of other animals (when we use visual intelligence as a proxy for general intelligence) and miss out on smart uses of other senses entirely. Rats are well-known for thwarting maze studies using things like fine sensitivity to slope, directional orientation using smell gradients across a room, or detecting the direction of researchers outside the maze based on micro-vibrations.
(Good book on the general topic of measuring animal intelligence: "Are We Smart Enough to Know How Smart Animals Are?" by Frans de Waal)
Anecdotally: Cats like fruit ninja.
Funny enough, so does my green cheek conure (small parrot about the side of a fist). Their beaks are made from keratin, like our nails, so it's conductive when touching the screen.
The hiss of the bombs gets him a bit angry though. Parrots hiss and it kind of sounds like that.
As an evolutionary cousin of a rat, the half-second delay after firing my weapon and the response would make me want to eat my whiskers.
This would completely kill any potential reward (and replace it with the opposite, frustration) you're trying to train me with, please fix immediately.
Lol you are correct! At least training them becomes smoother the faster you administer reward. Learning happens at different timescales in the brain, and immediate feedback (about <300 ms) yields the most reliable neural updates.
One of the most important traits of intelligence is understanding and appreciating delayed reward.
I saw a gambler win the jackpot. He was really excited and started gathering up all the chips he'd won. Why he was so excited to win a bunch of plastic chips, I'll never know. What's so great about plastic chips? Why was his brain so excited when all he was doing was gathering plastic chips? ...
A half-second delay doesn't mean your brain can't learn to make the precursor feel good.
The colorful plastic chips work well enough to trick our dumb monkey brains, just like red notification bubbles.
You missed the point. The chips are exchanged for money shortly after. The brain knows this.
Similar to the rat knowing the sugar comes very shortly after the task.
I think that’s the same idea behind clicker training for dogs. There’s a delay in giving them the actual treat, but the instantaneous click sound lets them now they did the thing that results in a treat
pavlov conditioning
not pavlov conditioning -- in Skinner's three term contingency, the stimulus context acquires meaning/significance related to the consequences of a response. a neutral or even negative stimulus (context) can become it's own reward through this process. this conditioned stimulus explains most animal and human experience. Humans are especially prone to constructing meaning based on the primal.
Think of the senses: sound becomes talking, music, etc. food become cuisine, obesity, and anorexia. eyes becomes art, movies, etc. desire becomes porn, s@m, etc.
meaning is constructed, socially constructed, or what skinner call "learning." His masterwork, long forgotten, is the "generic nature of stimulus and response." Generic as it open to near total manipulation
skinner was the man
Yep. And money is just another kind of chip. What's so great about money?
It gets exchanged for hamburgers
How can he not include a video of it working? D:
i think its just people being out of touch with reality. perhaps engineering minds not thinking enough product. its too commonplace for me to even complain about. someone builds something primarily visual - a robot, a GUI application, etc. and links to their github/etc but they ensure that there are no visuals included.
This is far from primarily visual, but I do understand your point. I could not take videos unfortunately. I mentioned it in other places, but this project took so long, I just wanted to put it out there and get some feedback. I'm glad that people are this receptive to it, and I hope someone would take the project over!
You built all this and saw a mouse rolling on a ball playing Doom, but didn't record it?
> I could not take videos unfortunately
Why?
At least there are 3d models of the dispositive and some pictures !
https://www.youtube.com/watch?v=y0wjaeEiin8&t=17s
That's the previous setup from four years ago, where the rats just run down a straight corridor.
This is the new YouTube, though not much on the account
https://www.youtube.com/shorts/mq2yfy23j7s
In the results section it looks like there's supposed to be an image of it but the link "placeholder_rat_playing.png" returns a 404 :(
because there's nothing to show
I'm opposed to this project because it involves using animals in medical experiments, which I believe is never ethically justifiable. It goes against basic moral and ethical principles regarding animal treatment. If the project were designed to allow animals to choose whether or not to participate, it would be more acceptable. Some scientists have already explored such approaches. By not giving animals a choice, you're limiting their freedom and potentially exposing them to physical or psychological harm through your simulation. As someone who advocates for animal rights, I'd prefer to see alternative methods that don't involve animals or allow them to participate voluntarily
You have interesting points but I feel you are jumping to conclusions. 'Curiosity not judgment' is a great rule for me, at least, because I rarely know enough to judge. Criticizing without knowing almost certainly makes me wrong, produces nothing, and costs me the opportunity to do something useful. It also wastes time and energy for multiple people.
It also is unfair to the experimenters and alienates them, when they could become allies and improve their methods. It alienates others; it makes you seem defensive and someone who lashes out unfairly - who wants to be involved with that? Even if the researchers agreed, would they want to have this judgmental, attacking person around?
For example, someone could ask: 'Hi - This is quite innovative. How are the animals introduced to the setup, trained, and experimented with? Are they basically required to play? What if they stop? Do they want to stop at the end of the session? Do they seek it out? Are there signs of stress or enjoyment? There is a bunch of innovation in animal research on giving them choices, and as we learn more about animal emotions and intelligence it makes much more sense to consider these things. This experiment seems like a perfect setup to explore some of those things; I'd love to engage with you on it, and/or here are some links to learn about it ...'. The researchers might love to help.
Maybe you know all that. I just hate to see good causes turned into alienation.
> Even if the researchers agreed, would they want to have this judgmental, attacking person around?
1. You calling the person above judgmental and attacking is not as tolerant either.
2. What about things that are morally wrong? Slave owners wouldn’t want to have a judgmental attacking person around either. Does that mean we have to have curious-discussions about slavery?
Your comment finds ways to make it adversarial. What are you trying to build?
> What about things that are morally wrong?
That takes possibly the most certain path toward evil: I think what they do is morally wrong so I can act without morality toward them. It's the rationalization of many awful acts and people and ideologies. Just look at the worst of what religions do.
We should not think we are somehow above or exempt from that error, or so above sin generally that we can preach.
> I think what they do is morally wrong so I can act without morality toward them.
I never said that this is how we should treat immorality. That’s somehow your interpretation of what I wrote.
> We should not think we are somehow above or exempt from that error, or so above sin generally that we can preach.
Preach.
[dead]
Given that we're not great at communicating with animals, what level of reward would be considered justifiable for participating in the study?
As in they could get a reward for starting the study...
Also, isn't this sort of "voluntary" testing a little unethical in itself? For example, testing an addictive drug on a rat, they don't know the downstream consequences since there is limited communication, but the immediate effects might be incredibly gratifying. It would lead to high "volunteer" rates but still expose them to massive harm.
From the author's comment up thread:
"when they learnt to run on the ball and how that influences their reward, they got hooked. I believe they enjoy not just the reward, they get a sense of how their actions influence the game and they like that. They would run on the ball so much at some point they wouldn't even bother drinking all the juice and it was just dripping on the setup."
Then make your own world where you don't benefit from the knowledge gained by means you don't approve of. This world is not that world.
I agree with you, but I would add that even so, it is still far better than 99.99% of scientific animal experiments, in which animals are usually subjected to severe suffering (for example, through the injection of human tumors) and are almost always killed at the end, as their tissues have to be examined.
We don't even give human beings that choice. Everyone gets to be part of the rat race, whether they like it or not.
I struggle to see how it's a acceptance that it's done at someone's home without the animal experiment safe guards and ethics oversight you'd get at a research facility.
Topped off with a plug to get your email address for their new startup.
Do your pets give you consent for everything you do with them?
I rolled my eyes at this at first but after looking at the setup they have I have to agree with you… there’s something viscerally horrifying about hooking something sentient up to a trackball with a VR “headset” with no way out.
What about work horses?
It’s something about messing with reality. Obviously I can’t know this since (as far as I know) I am not a rat, but I have to believe it’s profoundly disorienting for their little rat brains to interact with VR. At least a work horse can trust its senses.
Thank you, it’s good to read a thoughtful response instead of the median “ooh I love this” “i love rats and I love doom too”.
Human self-centredness is often insufferable.
>We reached the point of rat habituation but didn’t start training. Our rats (Todd, Kojima, Gabe) aged out before full testing. The setup works, but behavioral validation is pending.
Ah man, what a pity. That VR rig is awesome, but it doesn't really seem to me they are planning to continue these experiments, or do they?
How could you go to this amount of work and not continue??? Probably just waiting for other people to get mad at the lack of progress so that they attempt this themselves, and then he can challenge them to a deathmatch with his highly trainer ringer rats
When rats gaming championships become a thing, I will have the top team for sure! Seriously though, this project took way too much time. We built 2 versions of the setup, lots of hardware, software building, testing. I hope a behavioral lab or hardware enthusiasts would take it over and scale it. I did learn a lot in the process and would gladly support anyone who would continue it. In fact, I have a setup now laying around in my living room up for the taking :D
Seriously impressive work, and I do hope someone runs with the torch, it's always fascinating to see what our little rodent buddies are capable of.
Even if you're sort of like an alien abducting them for experiments.... if you offered me a sugar drip to play doom in VR, I'd be there
They also seemed to basically do it for two weeks and then stop because the rats "aged out" - despite it supposedly working well and achieving concrete results.
I know rats aren't long-lived, but I would be interested to know how they determined the rats 'aged out'.
Could also be a complete failure they spent considerable effort in with zero results, and are hand waving and constructing a way to quit while claiming success.
I could see the rats not really connecting things and just puttering around. It's a pretty involved setup and I poor uptake on the part of the rats would be a steep disappointment.
I wonder if the team at id considered this when they released Doom: In 30 years rats will be forced to play it in exchange for sugar water.
I don't think they considered it, but I'm positive they would have found it absolutely hilarious
The year is 2034. Countless attempts at re-producing the sophisticated wetware of the brain have failed. Modeling research has proved unfruitful, with the curse of dimensionality afflicting every attempt at breaking the walls of general intelligence. With only a few million of capital left, and facing bankruptcy, they knew that only one option remained.
"Bring me the rats."
Douglas Adams would point out that this is just why the rats already trained us to play DOOM.
The mice, actually; the rats are never mentioned.
That I remember, but I have to work with the material I am given …
finally somebody gets it.
This is the coolest thing Ive seen in a long time. I wonder if we could use this to train squirrels to not get hit by cars.
That might be easier than training people to slow down and take caution when driving.
It absolutely would be easier
I think there would be more money saved by making sure deer don't get hit by cars...
Such technology already exists. https://en.wikipedia.org/wiki/Wildlife_crossing
Im talking about squirrels and modern suburbia. That's a protected crossing
It's not about the money, it's about the squirrels!
I think there is probably an anthropocentric flaw in the rat display.
Human eyes are side-by-side and forward, with a big binocular overlap and a clear vanishing point. A forward-facing curved screen fits that geometry well.
Rats' eyes, by contrast, are lateral. They have a much wider field of view, a tiny binocular zone, and use motion and contrast more than neat perspective lines. A single human-style "cinema screen" isn't laid out for a rat's optics or brain.
Perhaps if the scenes were rendered with a much wider, 250° FOV, it would help the rat understand what it was seeing better.
Or even rendered with two virtual cameras offset and angled apart, then stitch their outputs into one extra-wide view wrapped onto the curved display. That would approximate the rat’s much wider horizontal field of view and reduce the mismatch between where its eyes are actually looking and where the important visual information appears.
There are other differences in perception of color and motion, but fixing the FOV would be an immediate and relatively easy software fix.
Also in the genre of animals playing video games: ‘Pigs can play video games with their snouts, scientists find’ https://www.bbc.com/news/technology-56023720
Well, the most cyberpunk thing I have read today. No video of rats killing Daemons though, only a promising but broken link.
bUt liNk wOrks On mY cOmpUteR (seriously its not broken) ..but if so, check out this thread, tldr of the whole rodent VR setup: https://x.com/yolorun_capital/status/1996632980903620886?s=2...
To be clear, I can read the article just fine, but there are no daemon-killing mice videos in the article, only a potentially interesting broken link.
>Our rats (Todd, Kojima, Gabe) aged out before full testing.
A Carmack might've kept the system stable long enough
They already tempted fate by having Gabe as the third rat.
While everyone training AI, this man train a rat. Are you gonna release the open weights (or the rat)?
Great project btw!
Treat those rats well, now that you've trained them for combat.
A huge shortcoming with this setup is that rats don't just see in front of them, they see all around them. The screen right in front of his nose doesn't really simulate his environment the way it does for a human. Luckily doom has command line options for screens on the left and the right. But I imagine a lot of his awareness comes from what's above him.
Doom aside (very impressive), I love the concept of putting rats on a roller ball instead of in a wheel for exercise. It would be better on their backs.
In the year 2032, the rodrone wars broke out. What started as innocent video game experiments, had taken a dark turn. After the rats had perfected playing the classic Doom games in 2026, they were easily trained on more complex simulations. In 2028 the first rodent controlled surveillance drones were tested. The year after weaponised. And the first real deployments in warfare in 2030. They easily outperformed OpenAIs latest battle systems. Home robots were soon after remote controlled by rats, known as Rodots. Rodent intelligence escalated quickly as lab selections bred only the smartest specimens. It was only a matter of time before the takeover. Now rodots were building steadily more capable drones and bots. Long before humans could foresee the need for a defense against the rats, the rodrone attack on all humankind, was a fact...
Pets.com buys 8M sq feet of datacenter realestate in a deal with Oracle's liquidators, which is reported to include fitting the racks with hamster wheels and feeders. Sets sights on 400B IPO
> rodent controlled surveillance drones
See also: WWII era, pigeon-controlled guided bomb: https://en.wikipedia.org/wiki/Project_Pigeon
It's kind of amazing how similar this looks to a joke system my brother came up with in 2005: https://paradies.jeena.net/gourmetica-mentiri/2005/11/sysham...
checks notes I'm sorry, this doesn't qualify to run in the category "running DOOM on <device>"
Yet. 2035: running Doom on rat brains.
Growing Living Rat Neurons To Play... DOOM? | Part 1
https://www.youtube.com/watch?v=bEXefdbQDjw
Thanks! Just yesterday I was thinking it's high time for me to check if there's any progress on the Torrent Nexus, and here we are
Would fit right in with LLMs' recent HN front page predictions.
running rats on doom?
Here is a thought: Maybe we are all just living in some gigantic experiment, that some incomprehensibly advanced civilization has created, to make us indirectly play DOOM, without us even realizing. Any progress we make is just one step further in their experiment, a reward, every issue like war, climate change, corruption, put in place by them as a challenge, while behind it all, in higher dimension, it is translated all into input for their version of DOOM, while they are watching us perform for them. One of them found it funny or interesting, to give us our own DOOM to play with, and now their scientists are speculating, whether it is always the case, that when you give a test population a DOOM, that they will pass a DOOM onto other beings, recursively.
This is a very popular theme, unfortunately without any practical conclusion.
I believe the practical conclusion is right there and pretty simple: if so, as long as we can't break out of this open world simulation and/or decisively prove we are in one, the best we can do is aim for success in the simulation. Just like we should do in the non-simulated world.
So, practically, we should behave the same.
Read the Bible. Conclusion seems very clear: We need strong men to fight invading demons, or we are doomed.
Good news: Here's a cute video of Todd being habituated to the setup: https://x.com/yolorun_capital/status/1999598643339227564
Bad news: no video of them playing on this setup, just on the previous version. We iterated on v2 too long, our pet rats grew old and couldn't be trained. We open-sourced the hardware and software so others can build upon it. You can TLDR the whole thing in this thread: https://x.com/yolorun_capital/status/1996632980903620886?s=2...
...Also, here's my personal X, dm me if you have any questions, or would want to build it for a lab or for yourself: https://x.com/viktor_thoth
I was worried this was going to be one of the petri dish of neurons type things, but the reality of it was awesome. More gaming rigs for rats need to be developed - it looked like Todd was actually enjoying it, and you toned down the potentially scary bits in a good way.
Kudos for the experiment and giving your pet an awesome enriching environment!
From now on every future Doom game will have to contain a "rat" difficulty.
How can you go through this immense effort and not have and show a video of the rats playing?
<<< just to verify, yes, he is The Rat Coach. Reach him out if you want to build something cool. (first post we made on the project: https://news.ycombinator.com/item?id=46151150 )
You got pwned by: COMMON FIELD MOUSE
That will feel next level bad when it happens.
I did not expect this to be as thoughtful as it is. Kudos to the thought put into the setup and training. Building custom hardware for the rat is amazing.. I need something like this for my own setup.
Thank you, we really appreciate it! <3
For those afraid to click: This does not appear to hurt the rats or involve doing any surgery on them.
Is it ethical to turn rats into gamers?
It was tested on humans first
I think it might be _unethical_ to not spread the joy of playing Doom for the first time? Though, I’m not entirely sure there’s been enough research done about the effects of violent video games in rat gamer populations.
We've been putting rats in skinner boxes for a lot longer than we've been subjecting human gamers to them. I'd be more worried about the health effects of all that sugar water.
10 microliter doses though. Pretty teeny tiny. But it is a valid concern.
Giving them slightly acidic water off-rig and normal water while playing is another option - not the best either. I opted for sugar in the end, because they didn't spend much time on the setup; but this should be reconsidered if they on-rig for hours daily.
The trick is to save the treats till the end. They can learn to do a lot if they know they'll be rewarded later.
Check out Shadow The Rat on YouTube; she has a whole series on training them.
Gotta love Shadow!!!!
https://www.youtube.com/watch?v=AV9z0c1hjnA
It's more like locking humans up in the matrix. Note lines like "preventing it from leaving the apparatus" in the build guide. Would an ordinary gamer be restricted from exiting such a contraption?
I mean-- how do you think they made gaming mice? :)
Gamers: no, hot-tub streamers: yes
I first read doing any sugary to them. Well, that's nice.
I was very happy to see this. I’m fairly against live animal testing, but giving rats the joy of playing Doom??? I think I _may_ have to be OK with this.
“No animals were harmed in the making of this blog”.
Has anyone 3D printed the levels, VR headsets are one thing but IRL is above that.
Perhaps I’ve just done too much miniature wargaming.
Do the rats enjoy it and if so, will the rats teach other rats how to play?
This could give a whole new meaning to "the rat race"
Crazy insane, I love it !
Guys, you should do it with a cockroach ^^ https://makeagif.com/gif/fifth-element-remote-controlled-coc...
What about your first newborn? Can’t wait! xx
Unbelievable as it sounds, the US did this in the 1940’s, but with pigeons and actual bombs.
https://en.wikipedia.org/wiki/Project_Pigeon?wprov=sfti1#
I'm a little conflicted on the ethics of this ... but this would have made a kickass youtube channel.
Why? Unless you're against keeping pets in general, what's wrong with training your pet rat to play a video game?
Violent video gaming is especially good if it keeps them off social media. We know that's bad for them.
I haven't look deep into this, but all the first thing I could think of is Clockwork Orange.
Rat strapped into some contraption and forced to play a game.
I haven’t read/watched Clockwork Orange, but I can understand the dystopian image.
It’s just that, when I think about it, I don’t see a difference in kind between training the rat to play a video game and teaching a dog to roll over and play dead. Neither is natural animal behavior, and the animal doesn’t really have a choice in either case.
Very cool... but why not use something with higher res graphics, if you're going to go through the trouble of building the mechanical system. You could go ultra 4K and see if a rat begins to see... reality.
That's such a fun project. Some really excellent Show HN items this week, but this is my favourite.
thx. :3 years of hard work. you might check out yolorun.capital for similar projects.
Is this related to the Tank lab experiments at Princeton?
I’ve seen this before with the practical application being that you can keep the animal stationary while it thinks it’s moving, in order to record brain activity.
> Our rats (Todd, Kojima, Gabe)...
Missed opportunity to name them Neo, Morpheus and Trinity.
Wonderfully crazy. Should the rats have multiple weapons at their disposal, which do they prefer? How many demons escape their fate? Are there videos of gaming sessions? So many questions!
This website causes my browsers (Chrome/Brave) to use a huge amount of CPU. Firefox handles it better but still has higher CPU usage than I'd expect.
> It maximizes immersion without obstructing whisker space
hehe very important
although... if the walls moved and touched the whiskers in conjunction with the game could be something
yeah good point! we actually have air puffs for that. we extracted from DOOM when a wall is hit in-game. when that happens, the corresponding side puff activates.
that's cool, I assume the sound doesn't scare the rat/does its purpose, sounds like a better solution than a physical wall moving
also nice to see those classic cheap blue stepper motors ha
No gameplay footage?
Next, train squirrels to play Mirrors Edge.
How good are the rats and can we start using rats instead of bots? Get them gainfully employed.
That’s all well and good, but when are we going to be able to play DOOM on rats?
The headset "maximizes immersion without obstructing whisker space" - nice!
This isn't true of all rats
So true! A little over fifty percent of the rat population would rather watch fox news and spend more time oiling up those weapons for when "it hits the fan".
Grandpa rats shaking their paws in anger going "It's all of 'em darned video games spoiling this generation of rats! Back in my days..."
The rats should be included in the team mentions
>The impact of prolonged VR exposure on rat well-being needs further research.
Fairly benign, compared to all the other things that get done to rats in labs, I would think.
People are fsc#ing insane. And I love it !!!
Incoming site: Can It Play Doom?
Living to Win (rat version)
Based on the headline I’m disappointed it wasn’t multiple rats playing DOOM together
any vidjas of rats playing DOOM
no vidjas :(
Whats the fascination with Doom? I keep seeing running Doom on this and that and now Rats playing Doom. I mean ok, Doom runs on everything and now rats playing doom. So what?
Doom has some star power and is instantly recognizable to pretty much anyone, not just video game fans. Its engine is Free software that's easily adaptable to pretty much any input or output method. In this case Doom might also have made a better candidate than another first person shooter because you can't aim up or down in Doom, so there's no need to account for that in the rats' input.
Doom pioneered the FPS genre in the USA. So it has a cultural significance. I used to go to my friends house to play it. You deal with a whole range of aspects such as grapchics, sound, networking,.. Etc. So it's a fun challange for hackers to try to get it to work in the most strange of ways. In addition it's a simple goal (with a complex path). So it helps with the focus, and most importantly the fun programming challenges involved in making the game work.
one of the first widely known "3d" shooters. Doesn't require much in the way of hardware and there are lots of open source implementations , tools and mods for it. And it's a meme
Doom was a big milestone in video game history and 90s tech culture.
Why do videos go viral? Why do memes become memes?
well done!
After I drowned old T-14 HDD, I expected old DOS emulation on a local disk. Take Operation Neptune in single-player mode.
[1]: https://archive.ph/tCAGu
[2]: Live Run ISO: https://archive.org/details/msdos_Super_Solvers_Operation_Ne...
[dead]
Next: The stock market!
[dead]