Thanks for watching Alejandro! The PID tuning for this project was definitely tiresome, but in the end, it was fun and I am happy with how it turned out.
This is dope. I would like to use this for something completely different. Is there a way to get your build / code and test it out for our use case? Our use case involves LiDAR.
@@IGNACASLTU Possible to add a fail safe for this. Once a decision is made prioritize until success then move to next. There also is similar things human minds do as well. Choice paralysis is one and hesitation is another. You only need to make something that simulates a person and, people are not perfect.
A "Physical" Aimbot. That's So Novel~! Why bother playing cat and mouse with the developers, programming payloads that bypass anti-cheat systems, when one can instead, "Program" the Player? I'm Beyond Words at Your Brilliance. Bravo and Thank You for Sharing this idea with the World ✌️ 💕
I saw a similar ideas around 2013, but it was much more goofy compared to this. Someone used a metronome to click their mouse1 button in order to high alch items in osrs. I also remember someone in world of warcraft using a fan to click the spacebar to make their character jump in order to prevent the afk logout timer. But, engineering is engineering at the end of the day.
Completely random idea, if you lower the resolution of your monitor, you could make it take faster screenshots through python. I bet you’ve already tried this, but if not, it’s something to think about.
no clue. he knows exactly what hes doing too. The second this is mastered he'll make bank and ruin gaming as a whole. It'll be funny to watch the same people praising this get upset when no multiplayer games exist anymore bc of cheaters.
I'm not scared that cheaters in video games will have this technology, I'm scared that a school shooter will one day have this and strap it to a gun 0.0
Nice you should learn about programming a microcontroller like an Arduino, then high to control a motor with a motor driver with pwn, and then make sure everything is power well. Lastly how to use serial communication to talk from computer to your microcontroller.
hey Well done people like you should have more clicks and need to be more fameous. Keeep up the good work maybe some at some time in your Research you can create an aimbot who works in games like valorant and cs go
man this is awesome from a technical standpoint but it's sad people's gonna use it to ruin online games. If you eventually sell it, everyone dying is gonna be mad at you and your invention. Imagine if future games start to require a whole set of cameras in your room to detect your not using this stuff lmao
I absolutely do not plan on selling this. And I will never used this outside of custom or against AI. I am, scared of the ethical of the robot, but currently games cant handle the easy stuff let alone this.
Start simple first get an arduino and program that to drive around. Than look through open cv computer library in Python. Have the two systems talk to one another.
@@Kioma187 Thanks for the kind words, but it took a lot of learning to get to this point I am no genius. You can make something like this to with a little practice
can we just talk about the fact that this litteraly looks like a real/human run, it's flicky, it got travel time, and it's not just straight to the point movment/telportation...
No it doesn't Every single shot has a proportional overshoot based on the distance the mouse is moving. The motors are running too hard to the point that they have to correct themselves. Bots are detected by unnatual movement. I.E. recaptcha v2 "I'm not a robot" checkbox. It's literally recording how your mouse moves to see if it's bot-like. that paired with background details can be used to find most most bots without a second test. Runescape actually has a anti-bot script that detects this kind of movement as well.
the thing that gives it away is that it chooses the newly spawning circles too fast. like it can recognize the newly spawned circle in 1 frame. You can see it really nicely at 0:43
this reminds me of when they started to use computers for chess: it got a lot of hate but it teached us humans a lot of new things. keep up the good work man!
Very interesting insight Leo! That's kind of the mindset I have, but because of computers Chess has become a lot more theory driven because it. I understand peoples fear of what I am making, and it is validated.
@@KamalCarter Yo bro if I were you i wouldn't be scared or guilty or anything. To me, you are a pioneer pushing the boundaries of artificial intelligence technology and somebody who would go down in the history books. Even if the exact benefits of these discoveries for the human race are still not that clear, I believe that like a lot of breakthrough inventions of the past it will help us one way or another. And, if time comes somebody uses this tech for malicious intent, you are *not* to blame. You wouldn't blame Tim Berners Lee for all the bad people in the internet right now yeah? Keep up the good work man. You are making history.
@@KamalCarter osu! has customs skins so i think this wouldnt be a problem. However in osu you need to aim and click on specific timing as "Aproaching circle" aproaches circle what you need to click. Pixel scanning may be difficult to do. Easiest method is scanning beatmap file(it has all timings) but i dont know if it is what you want from this project. Some guy did neural network for osu - NeuroSama, that only aims without clicking (it played with Relax mode where you need only to aim). Hope it's helpful info
I am afraid! This technology can be used for evil and not for animal crossing gameplay. Great project can’t wait to see what else you can do in the future!
@@AdrianRO1918 so who cares this is actually really interesting and how it would be almost impossible to detect it i. fps games i just wanna know how far it can improve
you should fix your mouse little bit more secure in that frame. fixture should be rigid from all the way from the ground thru wheels, motors, acrylic plate and mouse.
heres a tip to make movement even faster. instead of directly trying to steer to the place, you might want to have it use spline motions to make sure the motors dont have to counter their momentum each time moving from left to right or right to left as much. it might just make the robot a bit faster.
It's real basic right now, pretty much a pixel bot but in hardware form. There are some dma cheats out there that are a step below this but far above it in terms of software.
Came looking for this after watching the "top" of the aimlabs leaderboards.. was no way they were human reactions.. it looked like something was reading the screen... didnt expect to find physical bots doing the work. Good job on the creation though, seriously impressive work.. though Skynet will make good use of that and this aint so good 😂
That can't work, there is an optical sensor on the bottom of the mouse, so it's movement is based on its relationship to the surface underneath it. You could remove friction and get the mouse to glide on the pad as it's moved, but I can't imagine that would be nearly reliable enough.
@@snake1776 I understood the comment as not moving the mouse, which has considerable mass therefore requiring more energy to move leading to acceleration etc, but instead moving something light weight under the mouse so that the sensor would then pick that up. The same principle with which ball mouses and trackballs work.
@@fakedeltatime perhaps you could keep the mouse holstered slightly above the mouse pad. The easiest way to do that would probably be having a platform attached to the mouse pad underneath it and then you could rig it to move in each axis. Although I don't know if that would be any better, moving the mousepad without it being attached to a platform would create issues
If the aim is to remove physical unwanted variables such as weight but still use physical movements as input, then using the old ball mouse (non-optic) would be the way to go. You just hook your servos to the rollers themselves. This makes the movements much accurate, and you remove friction and slippage from the mouse-to-mousepad contact. The inputs would still be based on physical movement and not purely software feeding mouse coordinates.
@@alternateaccount8105 I'm not sure of this honestly, those balls weren't at all perfect, there is a reason we changed them, but it definitely still could be preferable to moving it with a wooden box
Woah, cool stuff bro! Just to ask. There is no memory on your robot, like to remember which is the first and Second, Third and etc.. dot? Or it's coded to point at the closest dot?
My robot is always tracking where the dots on the screen are and updating, so no memory. And then from that list of points will go to the one closest to center.
@Kamal Carter pls stop my score in Aim Labs is 300 points less than that if you keep developing this and open source it I can already give up playin Competitive fps games
Would it be possible to instead of moving a mouse, the input be the same as the provided by the mouse? You could remove the sensor and mock it to make the mouse send those data to the computer.
Yes that is very possible to spoof the mouse input either by emulating the drivers or by software. However, good anti cheat can detect non human esque motion like that though machine learning. Lastly, it so much more fun to build robots than work all in software!
Not without better Software. You can play against cs bots that have half of the reaction time of pros and never miss. You can still win most 1v1s because the bots have no clue where to aim and need to do a 180 where you just need to click once.
why did noone think of anything like this a long time ago, instead of fighting the Anti-Cheat just let it do its thing and make the Aimbot a real life machine that controls your mouse its such a simple idea why noone thought abiut this a long time ago, we may never know
It's been done before, I believe I've seen some smart phone robots that actuate a stylus to play mobile games via webcam. It hasn't been a breakthrough in actual FPS games for a few reasons. 1.) It's not easy to build the robot part of this yourself. It's a DIY project in itself. 2.) The AI required to do this needs to be FAST. And until recently image recognition software was very slow. 3.) You do not need a physical machine to do this undetected. It's actually quite simple to hide simulated mouse/keyboard controls through any kind of device driver. So this is kind of a solution to a problem that does not exist. The mechanical portion of this actually causes input lag and therefore will just make the process slower. The one benefit here is that because the movements are not simulated, they appear smoother to anyone watching the replay. Which is something that takes a LOT of tweaking and math to do in code. The hardest thing to do (and why this is probably only working on aimlab) is to be able to tell the computer when to shoot. It's simple enough for orbs that are a certain size/color on a pre-determined background... but it's a lot harder to do in a game with shadows, different character models, different character size/shape/color body parts, etc. in a very populated game world. That kind of computation can be done -- but getting it to work WITHOUT touching the game files is very difficult. For example, a tree limb in the distance could end up being the same shape of the arm of a character model... how do you tell the computer what the difference is? Once you answer that, then you are probably left with something that is too slow to be effective. This will eventually be solved though as hardware continues to get faster and faster and AI libraries are becoming increasingly more powerful and open. Don't get me wrong, this is still a cool project and obviously impressive. I have no doubt a lot of hard work went into this! This is not something just anyone could put together and do.
@@Chrisxantixemox yeah but the thing is its not really ai, its more of a software that scans the current frame and looks for human like objects, calculates the distance that it needs to aim at, and then calculate how much it needs to move the robot (and also move the mouse) to match the desired spot, i don't think its really an ai that constantly learns, i think it is more like a software of itself rather than a machine learning algorithm bullshit boiiii
try using the same technique as optical stabilators use for camera lenses in phones or physical hard drive heads - electromagnetic movement. Set the mouse speed/sensitivity to highest and create a moving surface under the laser sensor and move it with electromagnetic coils with some smart schematic. That should be the fastest and most precise way to do this. Probably quite difficult also but worth a shot.
Quick question do you have mouse acceleration turned off on your pc? that could be giving you some inaccuracies in the way that your machine learning works and the actual distance traveled by the mouse
As a game fanatic, software developer who worked with PID, I am impressed. I hope I can assist somehow in improving this to beat humans
Thanks for watching Alejandro! The PID tuning for this project was definitely tiresome, but in the end, it was fun and I am happy with how it turned out.
I can see humanity will be attacked by robots like in 2 years from now.
try making ur own or improving it would love to see it how far this can improve
This is dope. I would like to use this for something completely different. Is there a way to get your build / code and test it out for our use case? Our use case involves LiDAR.
Well it's got me beat
0:32 I love the pause between two dots as the robot struggles with decision overload
Wow that's a good catch, it happened so fast! The robot did weird things when two points had the same distance
@@KamalCarter I completely understand, I do the exact same thing just 10× slower.
Bro fr. It feels so... human
That shows that AI can't be perfect
@@IGNACASLTU Possible to add a fail safe for this. Once a decision is made prioritize until success then move to next. There also is similar things human minds do as well. Choice paralysis is one and hesitation is another. You only need to make something that simulates a person and, people are not perfect.
A "Physical" Aimbot. That's So Novel~! Why bother playing cat and mouse with the developers, programming payloads that bypass anti-cheat systems, when one can instead, "Program" the Player? I'm Beyond Words at Your Brilliance. Bravo and Thank You for Sharing this idea with the World ✌️ 💕
I saw a similar ideas around 2013, but it was much more goofy compared to this. Someone used a metronome to click their mouse1 button in order to high alch items in osrs. I also remember someone in world of warcraft using a fan to click the spacebar to make their character jump in order to prevent the afk logout timer. But, engineering is engineering at the end of the day.
WTF CRACKED AIMER
It's been 6 months, is there an update to this bot? I wanna see if it can reach new heights...
super interesting!
Completely random idea, if you lower the resolution of your monitor, you could make it take faster screenshots through python. I bet you’ve already tried this, but if not, it’s something to think about.
Yes! This was big reason why the code work the monitor resolution was 800x600
so this is who i have been playing in apex. Well played....robo tray
How are they looting
I feel like game developers will make a fps type of captcha in games like valorant beside there preexisting anticheat
i really dont understand why people are praising this guy for this
Same tbh
no clue. he knows exactly what hes doing too. The second this is mastered he'll make bank and ruin gaming as a whole. It'll be funny to watch the same people praising this get upset when no multiplayer games exist anymore bc of cheaters.
hes cracked
I wonder if the score would get better with a Gaming Mouse?
Only if it has RGB LEDs
you could also connect it to a $1000 razer gaming chair for even better results
I hope you will continue in the developement and will implement an AI and try it in a real shooting game
Maybe no ?
@@omicron7735 Is that a question or an answer?
@@thisfeatureisbad We don't need any more cheaters in games. So testing in real shooting game is bad idea
@@omicron7735 I mean, using them against the cheaters or other AIs separated, not with real players.
I'm not scared that cheaters in video games will have this technology, I'm scared that a school shooter will one day have this and strap it to a gun 0.0
nah, by that time we will also have counter measure AI for the malicious AI actors.
I have no experience in electronics but what I should learn to connect the physical components? I
Nice you should learn about programming a microcontroller like an Arduino, then high to control a motor with a motor driver with pwn, and then make sure everything is power well. Lastly how to use serial communication to talk from computer to your microcontroller.
@@KamalCarter you my inspiration
If I am brand new in robotics how much time it take me to be able to build this project
hey Well done people like you should have more clicks and need to be more fameous. Keeep up the good work maybe some at some time in your Research you can create an aimbot who works in games like valorant and cs go
Thanks for watching and the kind words!
Why would he ruin peoples fun like that
I’ve always been interested in how anti cheat and cheats work, but do not commercialize it
man this is awesome from a technical standpoint but it's sad people's gonna use it to ruin online games. If you eventually sell it, everyone dying is gonna be mad at you and your invention. Imagine if future games start to require a whole set of cameras in your room to detect your not using this stuff lmao
I absolutely do not plan on selling this. And I will never used this outside of custom or against AI. I am, scared of the ethical of the robot, but currently games cant handle the easy stuff let alone this.
Aaaand the software aim still 99.9
Its not an aimbot, its a bot who’s aiming
I think if youll put the fov highest possible. You less likely overmove bc the mouse dont have to move that much
Sensitivity is the same no matter the fov
@@susfinder5754 yea but distance to travel is not the same ergo, overdrive is less.. well i think
but can it play osu? lol.
imagine this robot with logitech g pro superlight
How to make this??
Start simple first get an arduino and program that to drive around. Than look through open cv computer library in Python. Have the two systems talk to one another.
@@KamalCarter UR SO GENIUS!
@@Kioma187 Thanks for the kind words, but it took a lot of learning to get to this point I am no genius. You can make something like this to with a little practice
@@KamalCarter this is what u make u genius man don't be shy!
@@KamalCarter Was the teensy running C or C++ fine for this?
can you make it faster?
Get a better mouse with no acceleration
you using trash mouse, so i 99% sure you didnt event turned off mouse acceleration in windows
Can you do a script for Cronus zen? Maybe? Think about it. 😉
+rep
can we just talk about the fact that this litteraly looks like a real/human run, it's flicky, it got travel time, and it's not just straight to the point movment/telportation...
It's surreal x.x what an Amazing Time We Live In
No it doesn't Every single shot has a proportional overshoot based on the distance the mouse is moving. The motors are running too hard to the point that they have to correct themselves. Bots are detected by unnatual movement. I.E. recaptcha v2 "I'm not a robot" checkbox. It's literally recording how your mouse moves to see if it's bot-like. that paired with background details can be used to find most most bots without a second test.
Runescape actually has a anti-bot script that detects this kind of movement as well.
@@icefire5555 Yo RuneScape bot tracking is next level I don't know how they do it they're crazy good at catching bots over the years!
the thing that gives it away is that it chooses the newly spawning circles too fast. like it can recognize the newly spawned circle in 1 frame. You can see it really nicely at 0:43
its pretty obviously not human though lol
this reminds me of when they started to use computers for chess: it got a lot of hate but it teached us humans a lot of new things. keep up the good work man!
Very interesting insight Leo! That's kind of the mindset I have, but because of computers Chess has become a lot more theory driven because it. I understand peoples fear of what I am making, and it is validated.
@@KamalCarter Yo bro if I were you i wouldn't be scared or guilty or anything. To me, you are a pioneer pushing the boundaries of artificial intelligence technology and somebody who would go down in the history books. Even if the exact benefits of these discoveries for the human race are still not that clear, I believe that like a lot of breakthrough inventions of the past it will help us one way or another. And, if time comes somebody uses this tech for malicious intent, you are *not* to blame. You wouldn't blame Tim Berners Lee for all the bad people in the internet right now yeah? Keep up the good work man. You are making history.
@@KamalCarter honestly, im all for what you're doing, *especially* if it slowly kills competitive fps lol
@@harshsrivastava9570 lol that's funny 🤣. And thanks for watching.
How does this teach us anything? Chess bots taught us new strategies, aimbots just ruin competitive matches.
Lol I'd love to see you use this on the rhythm game called "Osu!" - It'd make for some really interesting results I'm sure.
I did think about doing this on OSU, and I might do that in the future.
@@KamalCarter It has to learn how to Click drag which might not be a big of a deal.
Right?
@@LameSlime Yeah click drag wouldn't be a problem more so recognizing the different colors osu has
@@KamalCarter osu! has customs skins so i think this wouldnt be a problem. However in osu you need to aim and click on specific timing as "Aproaching circle" aproaches circle what you need to click. Pixel scanning may be difficult to do. Easiest method is scanning beatmap file(it has all timings) but i dont know if it is what you want from this project. Some guy did neural network for osu - NeuroSama, that only aims without clicking (it played with Relax mode where you need only to aim). Hope it's helpful info
I am afraid! This technology can be used for evil and not for animal crossing gameplay. Great project can’t wait to see what else you can do in the future!
why tf you keep supporting him if you are literally saying you against him
@@AdrianRO1918 “animal crossing gameplay”
@@theone614777 "great project cant wait to see.." also "im afraid" lol
@@AdrianRO1918 so who cares this is actually really interesting and how it would be almost impossible to detect it i. fps games i just wanna know how far it can improve
@@LOVE-wq4ku so you dont give afuk about all those games and their players, got it
you should fix your mouse little bit more secure in that frame. fixture should be rigid from all the way from the ground thru wheels, motors, acrylic plate and mouse.
Bro still Tenz faster than this Factttt.......
im 14 year old and i also make some projects like you. i am on a learning stage and i was just amazed by ur creation! nice!!
Nice thanks for watching. Keep learning! It's a slow process, but worth it at the end.
heres a tip to make movement even faster.
instead of directly trying to steer to the place, you might want to have it use spline motions to make sure the motors dont have to counter their momentum each time moving from left to right or right to left as much. it might just make the robot a bit faster.
What if he could make the device move the surface with a fixed mouse.
Jesus you're telling me TenZ topped this by another 20k??
Yeah Tenz on another level. I could have potentially caught him in score but that would have been a lot more effort.
@@KamalCarter better Mouse!
the world record is like 170k lol
Add some random noise on the controls to make it less likely to be detected as a cheat.
This can be used on military
This would work well with Cyberdyne Systems! ;)
really!?
I want this for cs please.
Gonna make som' kids cry.
*"Evident beyond a reasonable doubt"*
ur smart and i respect you but this project is just another step to the dooming of my beloved hobby.
Make this thing open source… so that people will not reinvent wheel.
When are you going to trian it with valorant enemy’s on targets instead
I'm working on something like that I'm close putting finishing touches together
I'd really like to see its capabilities in an FPS game. The ultimate "CampBot".
It's real basic right now, pretty much a pixel bot but in hardware form. There are some dma cheats out there that are a step below this but far above it in terms of software.
I read this in Arnold Schwarzenegger's voice.
Came looking for this after watching the "top" of the aimlabs leaderboards.. was no way they were human reactions.. it looked like something was reading the screen... didnt expect to find physical bots doing the work.
Good job on the creation though, seriously impressive work.. though Skynet will make good use of that and this aint so good 😂
@@mrwalker82 thanks for watching. There are some real people on the leaderboard though and have those fast reflexes 😯
what if you use those 3D printers arm that controls the nozzle and replace the nozzle with sensors so it will be light enough to move fast
What about 3D print a perfect platform for the mouse to be hold.
Gg boys comp has been fun
Awesome project! You might achieve faster speeds by not moving the mouse but the "mouse pad" instead ;)
That can't work, there is an optical sensor on the bottom of the mouse, so it's movement is based on its relationship to the surface underneath it. You could remove friction and get the mouse to glide on the pad as it's moved, but I can't imagine that would be nearly reliable enough.
@@snake1776 I understood the comment as not moving the mouse, which has considerable mass therefore requiring more energy to move leading to acceleration etc, but instead moving something light weight under the mouse so that the sensor would then pick that up. The same principle with which ball mouses and trackballs work.
@@fakedeltatime perhaps you could keep the mouse holstered slightly above the mouse pad. The easiest way to do that would probably be having a platform attached to the mouse pad underneath it and then you could rig it to move in each axis. Although I don't know if that would be any better, moving the mousepad without it being attached to a platform would create issues
If the aim is to remove physical unwanted variables such as weight but still use physical movements as input, then using the old ball mouse (non-optic) would be the way to go. You just hook your servos to the rollers themselves. This makes the movements much accurate, and you remove friction and slippage from the mouse-to-mousepad contact. The inputs would still be based on physical movement and not purely software feeding mouse coordinates.
@@alternateaccount8105 I'm not sure of this honestly, those balls weren't at all perfect, there is a reason we changed them, but it definitely still could be preferable to moving it with a wooden box
getting a better mouse with a better sensor and this device will work wonders LOL
at this point u cant differentiate b/w this and a pro
Woah, cool stuff bro! Just to ask. There is no memory on your robot, like to remember which is the first and Second, Third and etc.. dot? Or it's coded to point at the closest dot?
My robot is always tracking where the dots on the screen are and updating, so no memory. And then from that list of points will go to the one closest to center.
@Kamal Carter pls stop my score in Aim Labs is 300 points less than that if you keep developing this and open source it I can already give up playin Competitive fps games
Dam bro that's bad crazy high score! I believe you can beat the robot and then you can say you're better than a robot.
Would it be possible to instead of moving a mouse, the input be the same as the provided by the mouse? You could remove the sensor and mock it to make the mouse send those data to the computer.
Yes that is very possible to spoof the mouse input either by emulating the drivers or by software. However, good anti cheat can detect non human esque motion like that though machine learning. Lastly, it so much more fun to build robots than work all in software!
POV: You are playing PUBG and are from China
Would be interesting to see a similar mechanism used on an analogue stick of a controller
This is definitely on my to do.
it looks like its actually a human moving the mouse but they just have godly aim
Now give it a g pro wireless...
3d printer port anytime soon? 😋
This how we being motion sickness be like : 🤣😂
Dude can u use this in valorant
how does this click?
Software sadly 😮💨 we fixing that V2
這有啥卵用 一台自動炮機都比這有用
Let's make him play a game 🤣
Let's talk numbers...
almost good like me 😂
Shit like this ruins games
Wtf, it's a gamer robot
R.i.p fps games. If this becomes prevalent in the cheating community of games, we are screwed.
This is amazing. It has potential to be a physical machine that is capable of beating any human in an fps 1v1
Not without better Software. You can play against cs bots that have half of the reaction time of pros and never miss. You can still win most 1v1s because the bots have no clue where to aim and need to do a 180 where you just need to click once.
and thats a good thing?
@@guymanguy its just a dope technology
@@seal7224 you need to see the wider implications of this though.
this. is. crazy.
this is a real aimbot
is that carterOW
no rgb
How the fuck
why did noone think of anything like this a long time ago, instead of fighting the Anti-Cheat just let it do its thing and make the Aimbot a real life machine that controls your mouse
its such a simple idea why noone thought abiut this a long time ago, we may never know
That's what I was thinking too, but it made for a fun video and I got to be the first to do something
It's been done before, I believe I've seen some smart phone robots that actuate a stylus to play mobile games via webcam.
It hasn't been a breakthrough in actual FPS games for a few reasons.
1.) It's not easy to build the robot part of this yourself. It's a DIY project in itself.
2.) The AI required to do this needs to be FAST. And until recently image recognition software was very slow.
3.) You do not need a physical machine to do this undetected. It's actually quite simple to hide simulated mouse/keyboard controls through any kind of device driver. So this is kind of a solution to a problem that does not exist. The mechanical portion of this actually causes input lag and therefore will just make the process slower. The one benefit here is that because the movements are not simulated, they appear smoother to anyone watching the replay. Which is something that takes a LOT of tweaking and math to do in code.
The hardest thing to do (and why this is probably only working on aimlab) is to be able to tell the computer when to shoot. It's simple enough for orbs that are a certain size/color on a pre-determined background... but it's a lot harder to do in a game with shadows, different character models, different character size/shape/color body parts, etc. in a very populated game world. That kind of computation can be done -- but getting it to work WITHOUT touching the game files is very difficult. For example, a tree limb in the distance could end up being the same shape of the arm of a character model... how do you tell the computer what the difference is? Once you answer that, then you are probably left with something that is too slow to be effective. This will eventually be solved though as hardware continues to get faster and faster and AI libraries are becoming increasingly more powerful and open.
Don't get me wrong, this is still a cool project and obviously impressive. I have no doubt a lot of hard work went into this! This is not something just anyone could put together and do.
@@Chrisxantixemox yeah it was probably written in c++ for the speed and all of that, i think that if it was written in python it would be slow as hell
@@Blonder_Studio python is has some really respectable libraries, especially in the way of image recognition and AI that are open source.
@@Chrisxantixemox yeah but the thing is its not really ai, its more of a software that scans the current frame and looks for human like objects, calculates the distance that it needs to aim at, and then calculate how much it needs to move the robot (and also move the mouse) to match the desired spot, i don't think its really an ai that constantly learns, i think it is more like a software of itself rather than a machine learning algorithm bullshit boiiii
And after this video, the CIA got to his house and he mysteriously vanished from the face of earth...
Ai has gone too far or maybe just far enough :)
Use a XY carriage system like a 3D printer
I can build it if you'd like, i just don't know code that well
try using the same technique as optical stabilators use for camera lenses in phones or physical hard drive heads - electromagnetic movement. Set the mouse speed/sensitivity to highest and create a moving surface under the laser sensor and move it with electromagnetic coils with some smart schematic. That should be the fastest and most precise way to do this. Probably quite difficult also but worth a shot.
interesting
bro alexa do be cracked at aimlab FR!
All you need is simulate keyboard movement then you can play without actually playing
wsad simulator
now just use a camera to register the game rather than imput from the pc so the robot is fully seperate.
Quick question do you have mouse acceleration turned off on your pc? that could be giving you some inaccuracies in the way that your machine learning works and the actual distance traveled by the mouse
Yeah I've tried with that on and off I haven't see tooooo big of a difference, but will explore that further.
tf2 bots finna have a field day with this one xDDDDD
what sensitivity to put to have a.i.m like that)
Aim lab sensitivity was at 1.0 and the mouse DPI was 1000
@@KamalCarter did you choose a sensitivity at random? and how much in cm/360
you need a higher dpi to make it work better
Never seen anything like this dude! i love it...❤❤❤❤
scrolled down to read the comments immediately and the ding sounds scared the shit out of me
Lol yeah they might had been a bit too loud compared to my audio.
That's how terminator and skynet started.......
this is fascinating aswell as frigthening to see towards fps games
Why have u invented this
To suffer
Yo!!!!!!!!! Do you have a github?
I do, but I'm not making this project publicly available as I don't want to be accused of selling or distributing cheats
@@KamalCarter 99.999999% of the people who want to cheat will use some software alternative. I really hope you figure out how to share the code :)
what have you done...
👀👀
What a nice gaming mouse...
Can we see how this works in real game?