RL needs a whole company of engineers to make it work. For some reason it fails a lot at complex environments . I've tried all possible ways to teach it how to play a fighting game (NSUNS4) but i have failed every time .
@fajaranuar43334 күн бұрын
Hmm I'm not a Fortnite or Ai expert but maybe you can train the AI in the following: Spawn a square arena Spawn bots from a particular point. Place the AI. Reward - Small Ai aims the bots Ai Shoots the bots Ai damages bots Ai Build walls Ai certain distance from bots Reward - Large Ai kills bots Ai clears the stage Punish- Small Bots aims Ai Bots shoots Ai Bots too near Ai Punish - Large Ai is damaged Ai is killed Just an idea
@NasrAllah-Codesignn6 күн бұрын
You're on the right track-let’s keep it rolling!
@christopherlperezcruz150714 күн бұрын
Can't believe you spent so much time on this. Pure Shooting AI can't get done. It's like hitting the lottery but you are also a baby. Only a pseudo AI can be made with a bot carrying the AI. Rewards have to be immediate and moving around yields no reward. Neither does shooting.
@christopherlperezcruz150714 күн бұрын
If you want to make a cool AI, Tony Hawk pro skater.
@Mangalify17 күн бұрын
how to make games for moblie
@sebastianblackfyre19 күн бұрын
I don’t think I can do what this man did
@dgepho20 күн бұрын
That was great. Thank you.
@PNKnews26 күн бұрын
The game you made is actually pretty cool and that's the whole story here. I think It would have made it to the first page of the app store without publisher as well. I disagree that hypercasual games are bad and the comparison to the bad movies is wrong. People love them and play them even just to kill time. A simple game with basic graphics can be just as much fun as a super deep, beautiful game. The main thing is excitement that it gives to players.
@OMEGA_SBАй бұрын
❤
@DeathPixelАй бұрын
Nice video. I looked a bit at your code, especially the step function of your custom env. I couldn't find any delay between when the agent performs an action and taking a screenshot for the next observation. So the only time that passes inbetween is the time it takes for the function to take a new screenshot (or your sampling rate is synced with the game at 60Hz?) Did you try maybe to reduce the sampling rate of the agent? Maybe that helps to reduce data over the time dimension and lead to faster learning?
@willkwanАй бұрын
Yeah, the sampling rate is limited by the speed of the screenshot, the reward function, and model inference. In my first training runs with the chicken (not documented or in the vid) I hadn't optimized the performance of the screencapture/reward function yet and was getting <10 samples per second. Since it actually had to shoot and dodge in those training runs, my guess is that such low FPS actually makes it harder for AI to learn, but who knows. The final training run had about 25 samples per second.
@DeathPixelАй бұрын
@@willkwan yeah, i think with low fps you need some kind of synchronization and ensure the rewards are given for the right action and not async. it may still work if the world is static and the only thing you do is navigating/traversing the map. but as soon as you have moving objects that you need to track/target, i think it quickly becomes a nightmare. so either synchronization or high fps. but high fps i think leads to a lot of redundant data and then I think a larger frame stack size is needed to make the observation space more diverse. but i don't know, just a guess.
@ben_jamin01Ай бұрын
"no useful game ai tool developments in 4 years.. fine I'll do it myself..." Absolute player 🙏🙏
@FryingPanАй бұрын
THIS IS THE GREATEST AI FORTNITE TRAINING OF ALL TIME!!
@SamiCScАй бұрын
The only notification I get from youtube is when you upload.
@AwareOCEАй бұрын
As someone heavily interested in game based AI, this is fucking awesome! Cool as hell work, cant wait to see your future projects!
@user-wq6wz9qg5lАй бұрын
🔥🔥🔥🔥
@seamusgalivan3179Ай бұрын
Love to see it Will! Hope you continue rolling out videos!
@mehvixАй бұрын
start 2:38
@Gazzar19Ай бұрын
Training in real time might take to long to get enough data
@SanoKeiАй бұрын
Very cool, but why didn't you use DrEureka?
@willkwanАй бұрын
Didn't know about it, but with how long the training takes and the lack of parallel environments, I guess it would take a long time to try automated reward function tuning.
@BrandonCourtАй бұрын
I feel like what we call "AI" in recent years is misleading and over hyped because it's too attached to an expectation of being close to achieving sentient AI and it easily SEEMS that way on surface. I think this technological breakthrough of LLM is inherently a dead end in terms of future progression when looking at it as 'method' of AI. It's more like Mimicked Intelligence. We should be calling it MI.
@leoh2502Ай бұрын
I agree! The only difference between the nigh useless GPT-3 and the amazing ChatGPT is that the latter has also been trained specifically mimic a human in its outputs.
@NeostormXLMAX3 күн бұрын
Alparently people might have found the true source of our consciousness being microtubules in our brains performing quantum wave collapses, in other words no ai exist without quantum computing
@LupusMechanicusАй бұрын
bro game sucks ass, the building is zoomer cancer. Pick a good game next time.
@juanmacias5922Ай бұрын
Time to slap a transformer model on it, and see what happens. LOL
@awaz4160Ай бұрын
amazing
@MakeDataUsefulАй бұрын
Thanks for sharing your learning Will, I'm looking forward to checking out the repo.
@dlgamingyt12Ай бұрын
Play The Map Called Fun World 4
@neerajkulkarni6506Ай бұрын
nice to see youre back!
@davidfu8174Ай бұрын
blessing for the algo
@orionh5535Ай бұрын
Nice to see your still e-alive. Nice to see the thought process behind a persomal process. I feel like chat GPT killed alot pf research not related to having a chatbot as the final product behind AI research.
@BrandonCourtАй бұрын
Exactly. It's a dead end in terms of actually progressing AI.
@kushaagra098Ай бұрын
this video fs will go viral lol
@LakeMinnetonkaАй бұрын
Wtf I actually played super aquarium
@markdavies981Ай бұрын
I'm four years late with my comment, but this is a really interesting video. Your words contain a lot of wisdom. I'm working on my own mobile game, and my main motivation is to create something that I like to play myself. Like you, I've also downloaded a ton of simple games for motivation. Most of the games I've downloaded could be better. I play them once or twice, or sometimes only for a few seconds, and then delete them. In the meantime I've probably watched a couple of video ads, which is annoying as hell. I much prefer to play my own game because I like it, and I am addicted to trying to beat my previous high score. I'm not sure if that is a good sign or not, because I've no clue if anyone else will like it as much as I do. I guess I'll eventually find out if/when my game gets published.
@PersonWithDreamz27 күн бұрын
same here! im also working on a game to play for myself and for fun!
@mrqwakgamesАй бұрын
Reasons not to use a publisher #6231
@ItsAllGoodGamesАй бұрын
You know this was a real video, i sub
@dreammix9430Ай бұрын
Ha! I freaning love this boy!! Brilliant!
@CopperSteaker442 ай бұрын
•_•
@gabrielmarimba45932 ай бұрын
Keep going bro, youre Will reach greatness for sure
@edd72972 ай бұрын
Who else laughed when first seeing my boy? 😂😂
@PeterAdiSaputro3 ай бұрын
You're so cool because almost didn't blink in the video
@kekuKairinwakarimasen3 ай бұрын
What the fuck did he said chat? Chinese Incantations?
@jithukilladi75013 ай бұрын
His next video: How My Dumb Video Got 1.4M Views
@centripetal61573 ай бұрын
There seems to be a common mindset among software developers of chasing shiny and popular new trends in social media, or in general to create a digital product that will extract as much money as possible from consumers Logically this makes sense from a user / product demand point of view but its also a little depressing realizing software development ultimately boils down to selling popular digital products to consumer mass markets in an infinite everlasting grind
@ctwolf3 ай бұрын
Ohp. The software is working, algo just recommended this, Will, you killed it, you made it my g. Or you are currently making it more so. Keep it going. Excited to watch your meteoric rise.
@HereIsANewGuest3 ай бұрын
Serious fun here, very likable.
@amitch3 ай бұрын
What standing desk did you end up buying? And what was in shortlist?
@edench734 ай бұрын
what website is the best to train a generative ai model?
@tayyabnouman82404 ай бұрын
thanks for the video man. I wish I had discovered your channel earlier but hopefully its not too late
@sarerusoldone4 ай бұрын
nah bro, since all these games are only good at displaying ads, voodoo isn't a games publisher, they're an ads publisher lol
@richiedon91334 ай бұрын
creep
@LuisWeekes4 ай бұрын
is this possible to do on a normal laptop? and if not can you tell us how you used AWS?