Counter Strike as you've never seen it before
Вставка
- Опубліковано 15 тра 2024
- I went off the deep end and generated more AI images than you could eat. And they sure look tasty
0:00 - AI altered images
1:54 - How it works
3:24 - Style crossovers
4:15 - A Nuke Adventure
4:41 - Future of the tech
6:13 - Magical depth buffer
6:33 - Uses in CS:GO?
7:13 - HOW DOES IT KNOW SO MUCH
Thanks to Arcanumcsgo for sharing his setup and models used.
SO you want to make your own AI images?
Setting it up will take about half an hour:
1) Install everything needed for Stable Diffusion, AND Stable Diffusion • How to Install & Use S...
2) Install ControlNet (For being able to replicate existing images / CS:GO screenshots) • Stable Diffusion Contr...
3) Good, realistic models for environments and people:
civitai.com/models/4201/reali...
civitai.com/models/4823/delib...
4) Add positive / negative prompts to define what you want. Here's an example image, and what was entered to obtain it: civitai.com/gallery/64094?mod... - Ігри
You have no idea how fast I clicked on this after glancing at the title for 1 second only to realize this isn’t a source 2 video
Bro I jumped on this at work thinking it’s s2 dropping
Same and then I felt bad and watched the whole thing.
Same here. xD
yes idea
Click bait but I’ll allow it
for some reason i couldn't contain myself when he said "let's turn this place into mothafuckin lego" 4:25
It sounds so powerful coming from him
@@kampsudamien he rarely curses in videos, and when he says it with such emphasis, it hits you hard (but in a good way)
Literally the only time in YEARS i have ever heard him curse. And i've been following since 2010
Sounds too akward for me. "Madaffakin legoes"
@@loinen2298 I don't hear that at all
4:27 hearing philip swear is like hearing your childhood TV show character actors swearing off set for the first time
always impressed at your ability to keep coming up with new video ideas even when the game is in an update draught!
he truly is an amazing youtuber
tf2 fans be like
Best patch-notes reader ever.
you posted this 35 minutes ago, the video came out 36 minutes ago.. you watched this whole video in one minute? lol
@@UrsusHaitch Much better than WarOwl
I think this video only exists because you're an AI addict
Ahaha, you might be onto something here.
haha 420 likes, le funny
RIP Philip who died of heart attack hearing Valves announcement of CS2
😔
Warowl and 3kliksphilip doing AI CS videos on the same day?
Inconceivable!
An AI post from Warowl is a bit unusual, for philip it’s just an ordinary Thursday
@Noah Bailey Very true. Both were great though
Source 2 confirmed
Oh yeah, i noticed! Warowl also posted the video on the same day about AI
Are you all in Australia or something? Here in Brazil WarOwl Uploaded yesterday and 3kliks uploaded at midnight
The thought about how much AIs know has occurred to me before as well. It's insane how much knowledge is packed into those little bits of code. All of human history in image form, all stored in just a few gigs. You could spend an entire lifetime learning everything it knows without ever having to connect to the internet.
Not really. A vast amount of data is used to train them, but that data isn't actually documented within the AI itself.
I also wouldn't call it knowledge - it's just iterating from a very large data set. The generative process is more akin to dreaming than actual recollection or creativity. The AI knows nothing.
Bullshite man
@@givemeshrubs4295 can it do its magic offline or does the code need the internet to refer to the actual store of knowledge?
@@TheLegoJungle , it works offline; the AI training smushes multiple TB of source images and information into a 4GB file. The AI can't create any of its origin input images, but it can recognize the concepts in those images and tries its best to approximate what all the concepts you tell it to combine would look like.
@@TheLegoJungle like Mr or Mrs Azriel Satan said, the images themselves aren't actually stored in the AI. The images are just used to train it, so that it can recreate them at a later date.
As an analogy, you, as a human being, are the latest in a generation of other humans going back thousands of years. You are a product of all those humans.
But does that mean their full life history is encoded in your DNA? In order to walk, talk, etc, do you need to access a thousand years of genetic history? No. You were created by all these individuals, but you don't need access to their memory.
It's the same deal for AI - the training data *made* them, taught them how to turn verbal prompts into visual images, grew them - but now that the AIs are fully grown they don't need to refer to that information, in the same way that a human mind doesn't need to contain the record of millions of years of evolution to think.
That's just so much soul being put into simple things that is felt in your videos.
@@3kliksphilip AIs don't have souls, i hope ai doesn't go too far otherwise stuffs gona stop making sense?!?!
@@222toastedtoasters3 the man gives soul to the work, while the ai is a tool
Just wait till he realises Volvo have announced csgo source 2 cs2, bet he's on holiday again
@@xcodename11 ah yeah man, me too
@@222toastedtoasters3 Nothing has a soul. We’re all just electrical and chemical signals. AIs will soon have just as much “soul” as us.
typical valve, announcing big update when philips in vacation
I unironically love a lot of these images and would love to see more surrealistic CSGO map designs.
Someday it will definitely be able to do complex 3D, image how mindblowing it will be
4:27 that genuinely caught me by surprise lmao
The last few mintues really make you think about the wider implications of this tecnology. I don't know what the world was like before the internet, only bits and pieacs from media and my own early childhood, but I do remember the days right before it's mass adoption, when it went from a usful nichie to a nessesity that's part of everyday life. I have a sinking feeling AI is the next "internet" and my future children will be just as confused as I was when my parents told me about life before the World Wide Web.
"Back in my day people used to need to know how build things. Now AI just makes the blueprints" "Remember when people used to need to know how to run a nuclear powerplant? Now AI just gives us instructions or even automates it. Remember jobs in general?"
In the same way people don't know how to survive in the wild on their own and rely on civilization to survive, I really wonder if we're going to just become completely reliant on this thing that does everything for us
heresy
@@wondernayra who said I'm lamenting?
@@nawtmyrealnamelol"remember jobs in general" damn thats some positive thinking lol, best case scenario AI gets all the jobs you need a degree for, and humans are left to do hard labour since it would be too expensive to get and maintain machines for it
I’m a big nerd that loves these wacky AI videos from you. Big ups Philip!
I'm a big nerd that loves MF DOOM. Big ups T1ckin!
I enjoy theses videos, but as an artist and a programmer and some who can't actually afford AI art, these video makes me so fucking anxious.
I don't want to sound cheesy but I genuinely like your approach to things. your obsession with details is so satisfying to watch. you are one my favourite creators and I'll always follow your work. thanks for the details in the video description too! what a nice lad
Not to mention thanks for always adding subtitles to his videos!
Caught me off guard hearing Philip swear
Bro wake up sourse 2 is out!
People refreshing csgo twitter constantly while im refreshing 3kliksphilip constantly for him to cover cs2
Counterstrike 2 just got announced. Get on that video Phillip!
This would be an awesome cs quiz game where you have to guess based on the AI art alone which map is being shown
3kliks source 2 video waiting room
been refreshing for 12 hours straight
average holiday while cs2 is announced
The file size is the thing I nerd out about the most as well. It's insane to think all the words it knows can be stored in 2GB, let alone infinite potential images, words, and complex abstract relationships between them all.
It's scary how impressively good this looks.
Skynet is watching us
10:10 its almoust how we living beings store memories.
When we recall a city of "Brimingham, or Paris" we like to delude ourselves that our minds eye has perfect representation of the city, but in reality its a sorta jumbled mess like the things A.I. generates.
Actually, this technology works alot like an efficient living organism brain storage and memory recalling system
SOURCE TWO IS HERE PHILIP WE NEED YOU !
I really apreciate the way you use music in your videos, it makes them so unique and memorable, keep it up brother
DUDE STOP GOING ON VACATIONS
No, it's a tradition. Whenever Philip goes on vacation, there's a major CS update. We needed him to go on vacation 😁
Great video Philip, it's hard to understate just how well presented this was!
Didn't get lost in the reeds of technical jargon or oversimplifying concepts like most people that touch on AI images (especially ones that mention the impact for artists); instead, just a measured and articulate showcase, a well thought-out personal response/thoughts, and a glimpse into the trajectory of the world post-generative image AI - all without sugarcoating or fearmongering.
Artists will 100% be using these tools in the future, without question, regardless of the current ethical/legal dilemmas - these problems won't be around forever. The power of generative AI is astronomical, the scale of content that artists and creators can curate and shape is genuinely unprecedented, and absolutely key to the development of fuller and more immersive videogame worlds. We *right now* have the opportunity to try to push the development of these models in the right direction - to keep the benefits of these models in the hands of the public and common-good, not restricted to just those priveleged enough to have the money or influence to access paywalled models trained privately by mega-corporations. Open models are also much easier to identify the biases for, sheer exposure of millions of eyes tends to have that affect!
Anyway big tangent, great video. Makes me want to release my stable diffusion model that turns every photo into a CS:GO map :P
Pov: me still waiting for 3kliksphilip to make an indepth video on counter strike 2
"lets turn this place into motherfucking lego" got me
This video was absolutely gorgeous. The vast imagination I get from looking at these images and guided by the music I just feel like I'm dreaming. I love seeing the maps I've known for years and seeing them put in settings like space etc. Thank you for making this video Philip!
Amazing stuff as per usual. It always brightens my day when you upload and it always amazes me that you can take a subject that I might think is uninteresting and then make it something that I can't stop bothering my family about all day. I've learned about so much great stuff that I never would have even thought about because of you and I've had so many sad and boring days improved greatly just by seeing one of your videos in my notifications. Thank you :D
Thank you for going on a holiday 3kliksphilip! All of us owe you one!
I instinctively thought the video was going to end at 4:14. It has all the things that generally mark the ending of an xkliksphilip video.
ty for taking the time to put in stuff like the Deus Ex reference at 3:14
it really freshens up the video
Thank you for going on vacation Philip! We got CS2!!! You can come back now 😂
Imagine the possibilities with these depth maps.
You could hastily block together a scene in blender, just from a few simple shapes, then render it as a depth map, and give it to Stable Diffusion!
Or use a stereoscopic depth camera to create depth buffers of real life sceneries, and then reimagine them with the right text prompt
Fuck, that's genius.
That. Is. Wild.
Great video, as always! Loved the reference to Deus Ex with `why contain it?' 3:14
This video is amazing. It was so interesting to watch all the different concepts on the maps and the overall idea is great! I wouldn't mind watching more of these recreations even in other games!
always appreciate how in depth you go when making these videos
bhop
This will be a very strong tool for us mappers. We can get shots of a greybox map (no textures) then feed it a theme to generate super helpful concept art to guide textures/props as Philp said.
These videos are extremely interesting. Thanks for the full setup info In the description too
Fantastic video! My favorite across the kliksempire in a long time. You lure me in with some lighthearted img2img and you leave me pondering the future of creativity and questioning my understanding of computers. Amazing!
"Let's turn this place into mother-FUCKING Lego" Caught me off guard
That image of Inferno spawn completely fooled me. I didn't notice it was posted by you when I saw it on the reddit, and in the couple of minutes I stared at it, I was 100% convinced it was drawn by a person. It's just imperfect enough to not look like a screenshot with a filter, and the art style looks original and inspired. I don't want you to feel guilty at all, but it's crazy that all you really need is some sort of "master image" and some keywords for the AI to become an artist with their own style.
One of my favourite videos you made recently, thanks
The second half was lovely to listen to, your experiments and seeing the entire world in those 2 gigabytes, neat! :>
This video is really great. I have been thinking a lot about AI and what we can even use it for lately, needles to say I have been inspired. Keep making bangers Phillip
This is the best video I have seen by far! This AI image generation could give infinite ideas and inspiration concepts to mappers, level makers and level designers.
one of your best work yet! I LOL'd at that last painting
nice climax at the end! good job, phillll
Being an artist it depresses me how impressive ai generation is now
Is this how cobblers felt during the industrial revolution?
@@inn5268 I assume so... but perhaps they were a little gladder about the fact that everyone could then have high quality shoes. Artists don't generally seem glad that the average person can now make high quality art.
yeah, years of effort & skill being trumped by a machine in under a few seconds is certainly depressing ngl
remember that this is only the first stages of I.A, it is not even in his final form
@@thing4826 it's a big gray area (at least for me) whenever we have to explain why we dislike things like stable diffusing and other ai image generators in the field of arts. Also it could be since the program lifts from existing artworks from a lot of artists to create something new and we generally aren't kosher with that like we aren't kosher with tracers and claiming someone's elses art as their own. Anyway, I'm sure there's more qualified people in the business to explain it in better detail than I ever could, since I'm only just a part of this hellscape really.
Personally I think that it's okay to use them to create a concept quickly for visualization purposes, but anything outside of that, like generating one and then 'improving over it' then calling it their own, or using it to 'improve' an already existing artwork, or even monetizing it entirely, seems kind of off to me.
Inject a 1h video of imaginary CSGO maps with Philip's music straight into my veins
This is so amazing and fascinating I can't even put it in words. I'm super hyped for what is to come. And your videos are still the best on UA-cam. And I watch a lot of them I have to admit...
I love this video, especially the parts at the end :)
That's an interesting point about the tensor map somehow compressing vast swathes of human-generated information into several gigabytes.
I think the amazing thing about AI generated imagery and responses to prompts is that it doesn't actually "know" how to do anything. It makes arbitrary connections between different pieces of training data that even the programmers of the software don't know. They just feed it input and tell it the expected output and it works. Its going to change how people do things from now on because you no longer need to know how to do everything, you just need to know what you want as an end product.
@@3kliksphilip Yeah I guess we will have to see to what extent they will be trained on copyrighted images. I'm kind of praying it rules in favor of these AI companies because the images they create are insane.
I'm always grateful for watching your content, I love this kind of videos and the way you explain stuff in general. Thanks a lot for your time and work, greetings from Argentina ❤
From very early on in the video, I thought of how it might be used for conceptualizing map details or artstyle, and am glad to see you had the same line of thought. This is interesting.
Great video as always, Philip! Also, a suggestion if you're planning to make another video on this AI: You could try using Reshade's depth buffer ingame which gives you a sharper black and white image than the approximated 2D ones which might result in the AI taking better advantage of the depth clarity to produce a better quality result, just a thought.
It's always fun to watch videos of Philip having an existential crisis over AI
Thanks for providing instructions in your description! :)
Alot of these look like video game concept art, especizlly that bit where you made a map in hammer, took a screenshot, and tried to get different pictures of that place with dufferent themes and presets. Awesome video as always.
Phillip either didn't get a cs2 beta key or he's making the video he's dreamed of for years. hopefully the latter! I'm happy to wait either way.
They seemingly added accessibility features for blind people in CS2 based on the website
Thanks for the informative description.
Super exciting stuff! I imagine a future where every time you load into a map, the style is completely different but still plays the same. It will likely be a revolutionary way of keeping existing games feeling fresh.
Thank you for taking a vacation, 3kliksphilip! Counter-Strike 2 coming soon 😎
WHERE'S THE CS2 UPDATE VIDEO PLZ PHIL
He probably didn't get access
@@thatgreencat7978 That'd be insane of valve to not give him access with everything he's done for the community :/
@@Gainjey warowl didn't get access until he tweeted about it
@@thatgreencat7978 I just checked his twitter and by the look of it he's on a holiday - he's making jokes about not having internet and hoping he isn't missing anything big hahaha
The amount of fun you are having with this is audible the entire video, I can only imagine you recorded it with a massive grin.
Why do I enjoy your videos so much?
That is, to me, much harder to comprehend than all the AI wizardry shown in this video
(congrats on the video, Philip, as always)
Stable Diffusion is in my opinion better than Midjourney and other image generations simply because of all the LORA models and embeddings u can use.
what does that mean
Wake up
The last chapter of this video may be my favorite segment from one of your videos ever. I haven't paid much attention to AI generation besides the occasional senseless dall-e mini prompt just to see what it would do; that segment really put into perspective the scale of this technology.
Would also love to see a map by you filled with AI generated textures to see what limitations it has.
What a time to be alive 😮 This is so amazing, tremendous content as always 🎉💪🏻
He's probably out on a holiday
LITTLE DID HE KNOW, IT WOULD BE 5 DAYS LATER!!!!!!!!!!
The title is really underselling this vid, it’s definitely a favorite. Even just visually it’s very impressive
This is fucking incredible, and your presentation of such a vast and complex technology in a digestible and entertaining way is equally incredible. Love your videos!
I love AI stuff, though it's a scary thought for artistry. Being an artist is already a low-paid and underemployed field. I imagine a future where a game designer, vfx artist etc. using AI to create art, and then personally brushing up the work. It will be streamlined and made quick, like with many other jobs that have been automated before.
As a job? Probably, people only look at the results and if AI makes art just as good as us, then we may become obsolete
But that goes for digital art right now, and they haven't created comics and animes... yet
But for actual paintings it's not worth to make a robot draw and paint on a Canvas like a real artist, and artists enjoy making the painting just as much as thr result, while AI just goes straight to the end, i do appreciate watching someone draw something, and explaining his thought process you can't really get that from an AI as it's all inside his code which is purposelly made to be quick to get the results as soon as possible
AI isn't some tool, it only spits out "finished" stuff. To manually brush up that and fix things would literally take longer than creating it fron scratch. AI doesn't automate or simplify the art creation process but instead the only goal is replacing artists. What a dystopian world we live in
CS2.0 ? Next CS? Hello there ? .... :D Im waiting ........ :D
WE ALL ARE :D HE DIED OF HEART ATTACK PROBABLY
Very cool! Love the de_inferno volcano renders
Seems like Philip has gone down the Diffusion rabbit hole.
Just came from cs2
Wow it's another AI generated video!
Edit: Wow, AI really is getting very impressive. We've only gotta make a skeleton and use it as input to get infinite beautiful results! o-o
Dude, this is your best video so far
Love your channel ❤
Why did you give me more of an existential crisis at the end than I did myself when seeing things like this for the first time
AI art shows what's possible when you combine the greatest minds in tech, with a total disregard of copyright, and ownership (and possibly ethics?). I imagine one day AI images will be viewed the same as piracy, considering the amount of infringement needed to make AI images look as good as they do.
Bro, they gonna listen to you now when you will say "Valve pls fix"! :D
Honestly sat open-mouthed during that last "how does it know so much" section. Absolutely insane.
I genuinely love this video, so much emotional depth and effort
beautiful
WAKE UP PHILIPS
CS2 ANNOUNCEMENT IS OUT
hope you're not on vaccation during counter strike 2 reveal!
omg cbble!!! I haven't seen it in YEARS! Also props to you Philip for covering AI stuff without the typical 'oh no, AI bad and dangerous' thing. Really refreshing!
I'm the guy from the Reddit post :) So awesome seeing some of the stuff I generated in one of your videos, the future of AI will be incredible. (And horrifying.. but incredible nonetheless.)
awesome
1:48 "two tunnels made of penises" 😂😂
SOURCE 2 COME ON
3:13 I inmediately knew where that quote was from, something clicked inside my head and now i'm downloading deus ex again
The possibilities from this kind of AI in model and texture generation are really cool for future game and map creation.