The big issue with the osis proof of concept is the persistance window For example if you look up at the sky, a few frames then look back down the world model has lost all object permance. That is not an easy issue to solve in terms of compute imo That is why I believe we are more likely to see style tranfer of raw render engine output used to elevate end user visuals sooner than we will see game built from a transformer model
Pf. This is just a drop in the water to the evidence surrounding that idea. Double slit experiment, relativity, evidence leaning more towards lack of freewill over actually having it, then you got religions that tie in a lot of these ideas, there's literal ancient stories of other realms where "gods" can enter at will (simulations??), create other beings, eternal recursive realities, and the whole notion of god being in all places at once ties into the block universe theory, or if its presentism (which is the theory that only the present exists) it can still tie into simulations as if we're just live rendering as we go.
I swear in a decade we might acctually have ai generated aaa games made for you specifically, people don't get how big this is, it's simulating minecraft with no game code, this is as bad as it will ever be.
Yet the game code is absolutely necessary to train AI before it can "dream minecraft". Without a working reality (in this case, Minecraft), the AI cannot learn a thing! The key insight is that AI can reproduce the reality it is trained upon, just as our own minds recreate the reality of the waking world in our dreams. The hallucinations that plague AI are a feature of our own dreams as well.
@@jamesharford9788 There is no need for a baseline game. Minecraft (and all other games) are created mechanistically as an abstraction of our world by agents (us). Oasis is just a glimpse into the generative comparison of how our brains work. There are also many dimensions of the complexity gauge, higher level abstractions like language, give us LLMs which will be able to programmatically and methodically create simulations using formalized systems, and low level abstractions (oasis) can be used as building blocks following instructions of the higher level agents. At some point, everything will be a simulation (if this isn't one already)
Love how they called this AI program the same name as the software platform in "Ready Player One" which its interactive real-time virtual world is built on.
Reminds me of the comedy series Red Dwarf. They had a VR headset called 'better than life' which nearly wiped out humans as anyone using it didn't want to leave. Hope to see this before I die :)
heading towards that Star Trek Holo deck world where even the characters believe they are living out their lives in a real world. That simulation hypothosis seams less wacky every day.
if this was trained on game world state as well as visuals, it would naturally be able to interact with actual long term game state. Keep working from there, and you could move towards an AI that can work in any style and with any state, and then you just have a realtime game generator
No object permanence. The model should be learning how to generate the actual state and map of the game from the visuals not just generating some diffusion based image. Kinda silly.
Yes, this is amazing. I didn't even think we would be at this level this fast. If it was realtime coloring book generation it would be impressive. These are just the stepping stones to something jaw dropping and AI has just taken a giant first step
Another terrific upload, Wes! The process by which Oasis imagines the Minecraft world fascinates. I'm reminded of one bedtime years ago, when my brain imagined a lucid dream of fluid animation after I'd watched hours of awesome anime for some days. I trained my brain on anime until it unexpectedly imagined its own sequence. I'm sure others have experienced similar instances of spontaneous generation after mass consumption, in some domain or other. I'm bothered by the threat of unemployment this kind of AI poses to human game devs, though!
if we live in a simulation, i bet quantum physics not being deterministic is just the ai that simulates us not loading unnecessary data. kind of like not rendering all chunks of a minecraft world at all times.
Nice theory. Never thought about that, makes sense like the double slit experiment showing how particle reactions change if theres a spectator, tho with quantum physics it could just be kind of like the limit of how far we can see within this reality, to not render unnecessarily like you said.
Hey, I've been fascinated by the doom reproduction already, what you showed here with minecraft is already an enormous step forward. Although the interpretation of "playing" the video game seems much more like a dream as you said, for me it's close to the beginning of image generated content we saw with AI, first dreamt picture were assimilated as "hallucinations". There were crazy colors and parts of image only - but look where we are now. I can only see the potential in this technology getting crispier / sharper. Imagine us prompting how we want to design a game - gameplay and UI wise, the AI would then mix and match parts of its training data to provide the asked experience. It's advancing so fast it's kind of terrifying. Thx for the video.
I have mentioned in the past that "AI will replace software" and this is what I meant. There will be no code, just an AI. I understand that coders have a hard time with that, but I've been a coder for 30 years, and it is simply obvious to me. Next up will be, there will be no operating system, just an AI. Yes, hard to fit that into the Von Neumann architecture, but guess what ... there will be no architecture, either. I know, sounds nutty, but wait another year or two, and you'll see how this works.
Im about to graduate from uni next year - com science. It is totally clear to me that i wont ever be a junior developer like others were. Programming is incredibly difficult and im pretty much in awe of those of you that learned and mastered a domain but its clear that its difficulty and complexity is hindering our collective ability to utilise it. i can imagine that VS Code just becomes some kind of virtual interface that you talk to and collaborate on a piece of software with the Ai managing everything in the back end. This is also how people like Jensen Huang see programming so its obvious AND inevitable!
@@Lvxurie Definitely, its quite strange how in every field, even in fields that require so much knowledge and intelligence, that ego is so high and mighty.. it blinds us from seeing beyond what we can do in the moment. I like your point, it's such a complex task to program even with teams of people, its crazy to think what will be possible when AI gets good enough to do a lot of these things better than us, as it can utilize it's computational abilities, which of course we dont have.
I tried it out. The biggest obstacle to the current version being used as an engine is object impermanence. I think it needs metadata available relating to what's not currently visible. This isn't easy for training from videos, but I'd think would be relatively possible for synthetic training data. For example an "off screen" permanent inventory and 360 fish eye camera. For larger scale coherence I could imagine a form of RAG. If we could estimate the position and direction of view, you could throw in frames from earlier in the gameplay that look in that direction as extra context for what should be visible.
I see this as a dream machine for AGI to work out how the physical world works so training can be on going in lieu of upfront only. As an AGI system interacts with the real world every day it could take all those interactions and dream about all the other possible outcomes that didn't happen so to be trained how to react in the future. The AI would do this during downtime or offload the dream process to then integrate the dreams at a later time.
It’s amazing how fast this runs! I think the speed comes from using consistent seeds for each frame, so it’s not recreating everything from scratch, which helps keep things looking smooth and consistent. Plus, it seems like they might be using interpolation and temporal consistency, which help blend frames together naturally. And running on a TPU really boosts performance, especially with these optimized diffusion techniques.
❤The visuals or edit were pretty cool. I don’t know if you edited it, but it might be time-consuming to gather all the resources and put it into a timeline. I can’t wait for true multimodal AI, basically using computers and all the senses and working on our behalf but at top speed with accuracy. also whole video with each tangents/vectors were preety cool thank you soo very much for your contribution with every videos --- oasis model doesnt have memory(coordinates or map) as of now basically no coherence/cohesiveness of gameplay anytime you move new thing popsup weird stuffs like i wonder the bts of game cant wait the future evolution of gaming what it would look like & also would love to get any games recommendations
The next 5 years are gonna be absolute insanity, let alone the next 10. I for one can't wait for all the advancements everywhere, ethicality and morality dont concern me, as those are just human constructs that we will forever fight over whos "right" and "wrong" when in fact no one is either, objectively. Life will keep going no matter what happens, and im excited to see all that comes from it.
"Researchers made sure that landscapes remained consistent over time." And that was demonstrated in this video? I saw the world shift and change throughout the video.
This doesn't apply to just simulated worlds and games. You could also generate business applications and UIs on the fly instead of asking the agent to use an existing application (like Claude Computer Use). Which would be theoretically way easier than a full on simulated world. This is hinting to me as the end of coding all together and the beginning of full AI-generated experiences that are customized to whatever the user requests
This way you can create endless games, the script for which will be written on the fly by o1, that is, the quality of development is not inferior to ordinary games.
Because I’ve been gaming for decades to include environments like a EverQuest and World of Warcraft I think that it’s helped me develop the ability to create 3-D worlds in my mind when I need to think about something complex. This is a lot higher Fidelity than what my mind creates so yes, I think it’s very cool. I just wished that I didn’t have to use this against humanity because of all the crime but I really don’t have a choice.
They will absolutely use this to enslave humanity autonomously so that they don’t even have to manage any of you anymore and they can just run off and use all the resources
Wes, once they get all the data that they need to train the systems they’re going to get rid of us. Wars, disease, or anything else you can think of, but we will be gone if people don’t defend themselves. They’ve already been doing this through throughout humanity, but this time there will be nothing anyone can do once they have it in place because they’ll control all of the media and no matter how loud you yell you’ll never be heard. You will just disappear physically, and all of your data will be deleted like you never existed..
yeah.... i do think that we will use this tech for ... architecting and whatever you said... or mainly to generate the Brazzers videos we all dream off to play out the way we want with the faces of whoever we want.... i don't think you checked lately the stable diffusion models people have on Civit Ai.... anyway, yeah sure, if it makes you feel good, we will create virtual sets or architecture or whatever said, yeah yeah, definitely that's what it will be used for.
So, where are we headed in the end, Mr. Roth? Is this to teach ourselves that we are merely characters in a simulation of life, with no chance of peering beyond our universe? So many questions… in time, I suppose.
"Deep understanding" Bro if you look inside a color for 0.5 sec the world melts away, and if you walk backward you zoom out infinitely. Game can't even distinguish up/down. Also it is only recalling minecraft footage from it's training data. This won't lead to any "new" AI generated games, same as LLM's it will only do a mashup of it's training data, while not creating anything new.
You don't even understand what creation is.. lmao. Tf you think humans do? We're just conjuring up ideas out of thin air? No. We use our "training data" (knowledge) of experiences to solicit an outcome, therefore "mixing" whats in our brain to create something new. We're literally just getting started with AI models and we've gotten very far, but you think we won't go any further? Genius.
Yes, we're still early days with AI. Think back to cell phone developmental sophistication, and we're at the equivalent stage of just getting number button alpha-numeric texting. We haven't even gotten to Flip phone, color display, and are still quite a way away from an "iphone" moment. Looking at development this way, seeing where we are, and where we're going, and what could be in a thing even next year or the year after that, and then bigger "moon shot" realizations like AGI, "singularity", SAI, and so on, we're in interesting times. Further, unlike simply watching men walk on the moon in grainy black and white low resolution TV, AI is going to proliferate, and be in use almost everywhere. Some domains will be slow starters, but, like electricity, AI will creep into everyone's lives, and it'll be far more sophisticated than what we have now. In less than 10 years, Embodied AI with actually useful robots that do all the chores, housekeeping, upkeep, will be a thing in middle class homes. There WILL be mistakes along the way. Remember Tay, the early AI chatbot that when exposed to the wild wild internet was influenced into becoming a racist, sexist, nazi in a matter of hours? Mistakes will be made. Anyway, this minecraft bot is encouraging. The real-time modeling of whole worlds could act as predictive modeling overlay for real-world embodied AI to anticipate the real world around it, recognizing variables we may not even notice. This could also work as a function of next generation AR glasses in allowing the AR glasses wearer to see the world like never before, and even have a sort of "ESP" to predict real world environments for what can be expected. You can have Spock, or the android Data whispering in your ear that there's 76.2% probability that the thing is going to do a thing, and your logical course of action would be to do another other thing. I could be wrong in my predictions about these things. I have lots of predictions and anticipation regarding our near future in stages close, medium and farther out. I could be wrong about all of it, but, that's ok. Some of what I see, at least the good stuff will certainly be nice should it come. :)
Weird that noone makes movies with human actors making low-budget movie that is later only enhanced by AI. Or noone upgrades old some terrible B-class movies that look miserable, but have a good story, for example.
how would hackers still hack other than hallucination manipulation, also how would the backend for servers and multiplayer even work, will the transformers just understand that or will we have some kind of converter
eventually, AI will design a new architecture and the digital world will be rebuild (by AI). We will just be the administrators, communicating the functionality we want and thats it really. The AI will go out and find a suitable solution/s and just let you know when its done
@Lvxurie hopefully it's ai on top of all of this and not Sam altman, or a government.... too late on both of those though, Elon trusted the wrong person with the future of ai
@@Markus-r6g That's why open source exists. And why there's so much competition, it levels the playing field, even if one hits breakthrough, others follow shortly after. Besides if ASI were to actually be created, no one would have control over it, even those who created it.
@@sinnwalker depends, we aren't sure if no one would have control, that's just one theory it could just as easily be used for evil, btw the 3 people that are leading AI are closed source and government associated, open AI, USA, china, open source is a pipe dream, not even a seed in comparison, keep dreaming Mr glass half full and ignore the glass is completely empty and all is lost
Was this video AI edited? Really not a fan of this new video style. Complete with the one word captions popping up as you speak. It just feels like filler, rather than placing emphasis. And the weird background always there, with a small window showing content. Feels off.
It’s both very impressive and also underwhelming. As nifty as these videos are, the reality is there is no consistency in the world. A tree is there in one moment, then gone next time you look.
wes s gone again. its been nice seeing his real life image. it was even bigger and better on some vids. hope you come back soon. I almost can't watch these vids with no newsperson on anyones channel
This has zero application to science, for discovering how the world actually works. You should know better. So when you are implying that it might and then explicitly encourage us to express our opinions in the comments - that is probably just an engagement bait. I really dislike these tactics in your videos. You are succumbing to algorithmic drift, however you might try to rationalize your descent into employing of every dark pattern there is. Is such grift really worth it for you?
But you know, like we said about the Chinese, learning and copying is one thing - coming up with novel ideas (games) in this instance, is another. So yes while it might be fantastic - we are far from "new" games running on llms (also I think the ESG departments will have a heart attack looking at the co2 footprint from all the gamers in the world having to run this 😅🫣)
"far" is a bit pessimistic imo. We got here in the matter of just a few years, with compounding progress imagine what the next few years will look like.
the fact that it can be done just means it will get better and better
The big issue with the osis proof of concept is the persistance window
For example if you look up at the sky, a few frames then look back down the world model has lost all object permance.
That is not an easy issue to solve in terms of compute imo
That is why I believe we are more likely to see style tranfer of raw render engine output used to elevate end user visuals sooner than we will see game built from a transformer model
Now convince me we aren't living in a simulation.
Pf. This is just a drop in the water to the evidence surrounding that idea. Double slit experiment, relativity, evidence leaning more towards lack of freewill over actually having it, then you got religions that tie in a lot of these ideas, there's literal ancient stories of other realms where "gods" can enter at will (simulations??), create other beings, eternal recursive realities, and the whole notion of god being in all places at once ties into the block universe theory, or if its presentism (which is the theory that only the present exists) it can still tie into simulations as if we're just live rendering as we go.
We’re one out of an asymptote of trending towards infinity.
I am a simulation, you are real with real problems having to make real money. 😊🖖👍
I swear in a decade we might acctually have ai generated aaa games made for you specifically, people don't get how big this is, it's simulating minecraft with no game code, this is as bad as it will ever be.
Decade ?
I'd say in 2 years.
@Walexo45 im optimistic af, i could see it happening in 5 years, but not sooner, i praise your optimism.
Yet the game code is absolutely necessary to train AI before it can "dream minecraft". Without a working reality (in this case, Minecraft), the AI cannot learn a thing! The key insight is that AI can reproduce the reality it is trained upon, just as our own minds recreate the reality of the waking world in our dreams. The hallucinations that plague AI are a feature of our own dreams as well.
@@jamesharford9788
There is no need for a baseline game. Minecraft (and all other games) are created mechanistically as an abstraction of our world by agents (us). Oasis is just a glimpse into the generative comparison of how our brains work. There are also many dimensions of the complexity gauge, higher level abstractions like language, give us LLMs which will be able to programmatically and methodically create simulations using formalized systems, and low level abstractions (oasis) can be used as building blocks following instructions of the higher level agents. At some point, everything will be a simulation (if this isn't one already)
Minecraft looks like a** though; trashy retro graphics that were dated 40 years ago.
Love how they called this AI program the same name as the software platform in "Ready Player One" which its interactive real-time virtual world is built on.
How do you carve a statue of an elephant?
You cut away everything that does not look like an elephant.
solutions i love em!
th-cam.com/users/shortsivsgPF2HT0Q?si=1fi8q5JZK8cvHcWU
This man never sleeps! Wes, you're a machine yourself... thanks for all your hard work keeping an eye out for this stuff for us!
It's a labor of love 💕
You are making videos. So it's basically lovemaking. 😂
Reminds me of the comedy series Red Dwarf. They had a VR headset called 'better than life' which nearly wiped out humans as anyone using it didn't want to leave. Hope to see this before I die :)
Before you die inside one of those you mean? 😅
@@ub-relax6800 Exactly 😄
heading towards that Star Trek Holo deck world where even the characters believe they are living out their lives in a real world. That simulation hypothosis seams less wacky every day.
IMAGINE THAT AI BUT TRAINED WITH IRL FOOTAGE.
if this was trained on game world state as well as visuals, it would naturally be able to interact with actual long term game state. Keep working from there, and you could move towards an AI that can work in any style and with any state, and then you just have a realtime game generator
No object permanence. The model should be learning how to generate the actual state and map of the game from the visuals not just generating some diffusion based image. Kinda silly.
Yes, this is amazing. I didn't even think we would be at this level this fast. If it was realtime coloring book generation it would be impressive. These are just the stepping stones to something jaw dropping and AI has just taken a giant first step
Another terrific upload, Wes! The process by which Oasis imagines the Minecraft world fascinates. I'm reminded of one bedtime years ago, when my brain imagined a lucid dream of fluid animation after I'd watched hours of awesome anime for some days. I trained my brain on anime until it unexpectedly imagined its own sequence. I'm sure others have experienced similar instances of spontaneous generation after mass consumption, in some domain or other. I'm bothered by the threat of unemployment this kind of AI poses to human game devs, though!
It looks like someone who watches nonstop Minecraft videos (without actually playing it for himself) having a long nightmare about it.
100% true. Had lots of those.
Almost certainly the experience the AI is having: a nightmare it cannot awaken from. Sort of like real life in 2024.
if we live in a simulation, i bet quantum physics not being deterministic is just the ai that simulates us not loading unnecessary data. kind of like not rendering all chunks of a minecraft world at all times.
Nice theory. Never thought about that, makes sense like the double slit experiment showing how particle reactions change if theres a spectator, tho with quantum physics it could just be kind of like the limit of how far we can see within this reality, to not render unnecessarily like you said.
@@sinnwalker never thoughr of that, really? I thought that was the main idea/conclusion of the experiment. Interesting...
Great video! Minecraft video in the background is incredibly distracting though. 😂
Why are you making voice over videos? By now I'm invested in your facial expressions as you bring us news like this.
I'm invested in his facial expressions too. Isn't that weird? 😮
Hey, I've been fascinated by the doom reproduction already, what you showed here with minecraft is already an enormous step forward. Although the interpretation of "playing" the video game seems much more like a dream as you said, for me it's close to the beginning of image generated content we saw with AI, first dreamt picture were assimilated as "hallucinations". There were crazy colors and parts of image only - but look where we are now. I can only see the potential in this technology getting crispier / sharper. Imagine us prompting how we want to design a game - gameplay and UI wise, the AI would then mix and match parts of its training data to provide the asked experience. It's advancing so fast it's kind of terrifying. Thx for the video.
I have mentioned in the past that "AI will replace software" and this is what I meant. There will be no code, just an AI. I understand that coders have a hard time with that, but I've been a coder for 30 years, and it is simply obvious to me. Next up will be, there will be no operating system, just an AI. Yes, hard to fit that into the Von Neumann architecture, but guess what ... there will be no architecture, either. I know, sounds nutty, but wait another year or two, and you'll see how this works.
Im about to graduate from uni next year - com science. It is totally clear to me that i wont ever be a junior developer like others were. Programming is incredibly difficult and im pretty much in awe of those of you that learned and mastered a domain but its clear that its difficulty and complexity is hindering our collective ability to utilise it.
i can imagine that VS Code just becomes some kind of virtual interface that you talk to and collaborate on a piece of software with the Ai managing everything in the back end.
This is also how people like Jensen Huang see programming so its obvious AND inevitable!
@@Lvxurie Definitely, its quite strange how in every field, even in fields that require so much knowledge and intelligence, that ego is so high and mighty.. it blinds us from seeing beyond what we can do in the moment. I like your point, it's such a complex task to program even with teams of people, its crazy to think what will be possible when AI gets good enough to do a lot of these things better than us, as it can utilize it's computational abilities, which of course we dont have.
I tried it out. The biggest obstacle to the current version being used as an engine is object impermanence. I think it needs metadata available relating to what's not currently visible. This isn't easy for training from videos, but I'd think would be relatively possible for synthetic training data. For example an "off screen" permanent inventory and 360 fish eye camera. For larger scale coherence I could imagine a form of RAG. If we could estimate the position and direction of view, you could throw in frames from earlier in the gameplay that look in that direction as extra context for what should be visible.
...and what's the first thing we do after realizing it's a dream... we FLY 🙏
I see this as a dream machine for AGI to work out how the physical world works so training can be on going in lieu of upfront only. As an AGI system interacts with the real world every day it could take all those interactions and dream about all the other possible outcomes that didn't happen so to be trained how to react in the future. The AI would do this during downtime or offload the dream process to then integrate the dreams at a later time.
It’s amazing how fast this runs! I think the speed comes from using consistent seeds for each frame, so it’s not recreating everything from scratch, which helps keep things looking smooth and consistent. Plus, it seems like they might be using interpolation and temporal consistency, which help blend frames together naturally. And running on a TPU really boosts performance, especially with these optimized diffusion techniques.
I have a game irl. Just no game dev ability digitally - so I'm eagerly tracking this stuff. Hopefully the barrier to entry diminishes.
❤The visuals or edit were pretty cool. I don’t know if you edited it, but it might be time-consuming to gather all the resources and put it into a timeline. I can’t wait for true multimodal AI, basically using computers and all the senses and working on our behalf but at top speed with accuracy.
also whole video with each tangents/vectors were preety cool thank you soo very much for your contribution with every videos
---
oasis model doesnt have memory(coordinates or map) as of now basically no coherence/cohesiveness of gameplay anytime you move new thing popsup weird stuffs like i wonder the bts of game
cant wait the future evolution of gaming what it would look like & also would love to get any games recommendations
Wes, you're making a lot of good choices lately, bravo... you are um DIFFUSING into what exactly? ;) 🙏
That's what I thought when I first saw it. It looks like the model is dreaming Minecraft.
Make this your mantra.. What a world we live in. 😅
The next 5 years are gonna be absolute insanity, let alone the next 10. I for one can't wait for all the advancements everywhere, ethicality and morality dont concern me, as those are just human constructs that we will forever fight over whos "right" and "wrong" when in fact no one is either, objectively. Life will keep going no matter what happens, and im excited to see all that comes from it.
"Researchers made sure that landscapes remained consistent over time." And that was demonstrated in this video? I saw the world shift and change throughout the video.
This will create a ready player one metaverse, as we have a prompt to describe it
Did you see the AI Minecraft nether speeedrun? It's super interesting lol
There is no English Audio Track for this video on my Samsung TV.
This doesn't apply to just simulated worlds and games. You could also generate business applications and UIs on the fly instead of asking the agent to use an existing application (like Claude Computer Use). Which would be theoretically way easier than a full on simulated world. This is hinting to me as the end of coding all together and the beginning of full AI-generated experiences that are customized to whatever the user requests
They must join it with the source code ai.
This way you can create endless games, the script for which will be written on the fly by o1, that is, the quality of development is not inferior to ordinary games.
Can’t wait till the actual oasis becomes a reality
Because I’ve been gaming for decades to include environments like a EverQuest and World of Warcraft I think that it’s helped me develop the ability to create 3-D worlds in my mind when I need to think about something complex. This is a lot higher Fidelity than what my mind creates so yes, I think it’s very cool. I just wished that I didn’t have to use this against humanity because of all the crime but I really don’t have a choice.
They will absolutely use this to enslave humanity autonomously so that they don’t even have to manage any of you anymore and they can just run off and use all the resources
Wes, once they get all the data that they need to train the systems they’re going to get rid of us. Wars, disease, or anything else you can think of, but we will be gone if people don’t defend themselves. They’ve already been doing this through throughout humanity, but this time there will be nothing anyone can do once they have it in place because they’ll control all of the media and no matter how loud you yell you’ll never be heard. You will just disappear physically, and all of your data will be deleted like you never existed..
yeah.... i do think that we will use this tech for ... architecting and whatever you said... or mainly to generate the Brazzers videos we all dream off to play out the way we want with the faces of whoever we want.... i don't think you checked lately the stable diffusion models people have on Civit Ai.... anyway, yeah sure, if it makes you feel good, we will create virtual sets or architecture or whatever said, yeah yeah, definitely that's what it will be used for.
So, where are we headed in the end, Mr. Roth? Is this to teach ourselves that we are merely characters in a simulation of life, with no chance of peering beyond our universe? So many questions… in time, I suppose.
Second Wes Roth video that is dubbed in a foreign language, and I cannot select the original audio track. Is my TH-cam broken?
Looks like a fever dream.
haha cool psychedelic AI game
in 4k of this and VR? you don't need mushroom my friend
AIRIS = Minecraft AI… Mining a mind all its own. 🤖🖖🤖👍
Woo!!
"Deep understanding" Bro if you look inside a color for 0.5 sec the world melts away, and if you walk backward you zoom out infinitely. Game can't even distinguish up/down.
Also it is only recalling minecraft footage from it's training data.
This won't lead to any "new" AI generated games, same as LLM's it will only do a mashup of it's training data, while not creating anything new.
Inability to recreate training data with 100% precision will be indistinguishable from creativity. Just the way it is in humans.
You don't even understand what creation is.. lmao. Tf you think humans do? We're just conjuring up ideas out of thin air? No. We use our "training data" (knowledge) of experiences to solicit an outcome, therefore "mixing" whats in our brain to create something new. We're literally just getting started with AI models and we've gotten very far, but you think we won't go any further? Genius.
why would we need OS or even apps later on?
Wes Roth - the used car salesman of AI content
Yep!!!
i love all of you, internet people.
Yes, we're still early days with AI. Think back to cell phone developmental sophistication, and we're at the equivalent stage of just getting number button alpha-numeric texting. We haven't even gotten to Flip phone, color display, and are still quite a way away from an "iphone" moment.
Looking at development this way, seeing where we are, and where we're going, and what could be in a thing even next year or the year after that, and then bigger "moon shot" realizations like AGI, "singularity", SAI, and so on, we're in interesting times.
Further, unlike simply watching men walk on the moon in grainy black and white low resolution TV, AI is going to proliferate, and be in use almost everywhere. Some domains will be slow starters, but, like electricity, AI will creep into everyone's lives, and it'll be far more sophisticated than what we have now.
In less than 10 years, Embodied AI with actually useful robots that do all the chores, housekeeping, upkeep, will be a thing in middle class homes.
There WILL be mistakes along the way. Remember Tay, the early AI chatbot that when exposed to the wild wild internet was influenced into becoming a racist, sexist, nazi in a matter of hours? Mistakes will be made.
Anyway, this minecraft bot is encouraging. The real-time modeling of whole worlds could act as predictive modeling overlay for real-world embodied AI to anticipate the real world around it, recognizing variables we may not even notice. This could also work as a function of next generation AR glasses in allowing the AR glasses wearer to see the world like never before, and even have a sort of "ESP" to predict real world environments for what can be expected. You can have Spock, or the android Data whispering in your ear that there's 76.2% probability that the thing is going to do a thing, and your logical course of action would be to do another other thing.
I could be wrong in my predictions about these things. I have lots of predictions and anticipation regarding our near future in stages close, medium and farther out. I could be wrong about all of it, but, that's ok.
Some of what I see, at least the good stuff will certainly be nice should it come. :)
how do we combine worldsim technologies like Omni with this? what is the intersection?
Oasis is ridiculous, it's like gaming on shrooms. Absolutely nonsensical but funny.
Sure
its not a dream videogame, its a game video dream.
Credo che sia letteralmente IMPRESSIONANTE!
Weird that noone makes movies with human actors making low-budget movie that is later only enhanced by AI. Or noone upgrades old some terrible B-class movies that look miserable, but have a good story, for example.
Physical rendering and infinite time.
wes you are a great communicator but cmon dude oasis y completely unplayable did you even try it out? you are being bias be careful
how would hackers still hack other than hallucination manipulation, also how would the backend for servers and multiplayer even work, will the transformers just understand that or will we have some kind of converter
eventually, AI will design a new architecture and the digital world will be rebuild (by AI). We will just be the administrators, communicating the functionality we want and thats it really. The AI will go out and find a suitable solution/s and just let you know when its done
@Lvxurie hopefully it's ai on top of all of this and not Sam altman, or a government.... too late on both of those though, Elon trusted the wrong person with the future of ai
@@Markus-r6g That's why open source exists. And why there's so much competition, it levels the playing field, even if one hits breakthrough, others follow shortly after. Besides if ASI were to actually be created, no one would have control over it, even those who created it.
@@sinnwalker depends, we aren't sure if no one would have control, that's just one theory it could just as easily be used for evil, btw the 3 people that are leading AI are closed source and government associated, open AI, USA, china, open source is a pipe dream, not even a seed in comparison, keep dreaming Mr glass half full and ignore the glass is completely empty and all is lost
Was this video AI edited?
Really not a fan of this new video style. Complete with the one word captions popping up as you speak. It just feels like filler, rather than placing emphasis.
And the weird background always there, with a small window showing content.
Feels off.
if you can "simulate minecraft" why not you photorealistic game footage to train the AI and get a photorealistic game ?
Why is the video audio German generated ai voice ?
i made it to that one part
Structured because pixels go together well and not because physics concept. 😡
Looks like a weird android's dream. Not playable. No practical application.
First step towar 4k real style Ready Player One.... toward dytopia maybe 🎉
❤❤❤❤❤
Do Androids Dream of Electric Sheep?
They do now!
It’s both very impressive and also underwhelming. As nifty as these videos are, the reality is there is no consistency in the world. A tree is there in one moment, then gone next time you look.
wes s gone again. its been nice seeing his real life image. it was even bigger and better on some vids. hope you come back soon. I almost can't watch these vids with no newsperson on anyones channel
This has zero application to science, for discovering how the world actually works. You should know better. So when you are implying that it might and then explicitly encourage us to express our opinions in the comments - that is probably just an engagement bait. I really dislike these tactics in your videos. You are succumbing to algorithmic drift, however you might try to rationalize your descent into employing of every dark pattern there is. Is such grift really worth it for you?
🖤🔥
But you know, like we said about the Chinese, learning and copying is one thing - coming up with novel ideas (games) in this instance, is another.
So yes while it might be fantastic - we are far from "new" games running on llms (also I think the ESG departments will have a heart attack looking at the co2 footprint from all the gamers in the world having to run this 😅🫣)
"far" is a bit pessimistic imo. We got here in the matter of just a few years, with compounding progress imagine what the next few years will look like.