I'll be honest, I love watching your videos. Every time a video comes out, it's like a gift for me, I'm just like "Oh, Matt's gonna show me some cool new AI thing I didn't know about!". Watching your videos is like a tradition for me, I've been watching you for almost a year now, thank you for everything Matt, keep it up, you're very cool.
Surreal for me to hear! I appreciate the kind words :) I'm pretty deprecative of myself so often times it's hard for me to picture the fact that real people watch my vids, but when I get comments like this it makes my day!!!
I played *_Oasis_* and I think the best way to describe it is that it feels like when I'm dreaming and the entire reality changes when I look in another direction. It's really amazing and strange to see how similar AI seems to be when compared to the world of dreams.
Totally agree, Everything keeps changing! I think the fact it reminds us of dreams speaks to how these models really do to some small capacity mimic the brain, or at least possess similar side effects.
@@MattVidPro i love this interconnection between subliminal spaces(dreaming, psychodellics and deep meditation) and AI, its like we are building something that maybe can share the consciousness that we think its an human thing, maybe its just an intellingence thing. Keep up your videos, they make keeping up with ai news these clickbait times we are living in.
@@MattVidPro Yep. Generating images (or sequences of images, i.e. videos) is very similar to visual imagination. When the logic centers of the brain are suppressed as we sleep, we get dreams; when these AIs don't have enough of a logic center trained into them, we get also dreams 😁
In ChatGPT, you can actually pause the advanced chat mode, view the conversation, select search the web, voice input, and continue the conversation. It will carry the information from advanced chat into search and search back into the advanced chat. It's a bit of a runaround, but it's almost there and can be done (with a couple extra steps between)
Thanks for the info! I figured something like that would be possible. So weird when open AI has like 8 diff models trying to use the same interface I guess..
@@resumindo857 do you have the paid version of GPT or the free? Search is only available on the paid version until later, when they roll it out to everyone "in the coming months". If you do have paid, check for app updates and maybe the current version will allow you to. (It’s also available via the website to paid users) To find it, click the "+" button, as if to upload a file, but instead click the globe button to use Search.
Jensen Huang (CEO of Nvidia) said the future of gaming will be pixel by pixel generated on-device and will not require any individually created models, music files etc. Imagine this but its GTA and the file size is 50mb lol
Plus, escaping from npc cops and winning fist fights against npc opponents becomes easy because they're predictable. What if npcs could develop new strategies because they remember what the player did the last time?
50mb? Have you *seen* how big these capable models are? 😂 That said, memory and HDD/SDD space are constantly increasing, so maybe one day soon an optimized model that's capable enough to generate 4K frames of a game that are actually temporally coherent and logical will fit on some future consumer device.
The new Claude is great when it comes to coding. Ever since the first version of Chatgpt I've been trying to get it to write a specific python program to do a very niche job and up to this point they have all failed. But the latest Claude 3.5 nailed it. First time and even streamlined it when I asked for it. Amazing
When you pronounced scent "teleportation" as "tele-operation", I immediately wondered if this means you used ElevenLabs for your audio😄Jokes aside, thanks for the video. With so many updates and changes, it is great to have such videos because I can't keep track of it!
Perplexity is my default search option. I got a year's Pro subscription FREE when I purchased the Rabbit R1 (don't ask, I haven't even opened it -LOL!) I will renew the Pro subscription when the time comes.
So I was watching this video as I do all of Matt's videos, just chilling out, and then I thought I spied my boy (dog) in the corner, and then get mentioned. That was awesome. Also agree on closer look that Ideogram did a better job with the lemon monster. I hadn't noticed that weird little stormtrooper in the bottom right among the other things Matt pointed out. Very cool having so much competition in the image generation field now. and just getting better. Another great vid.
Oh how I HOPE you mention Oasis 500m... My God... It is interesting to have 100 cows turn in to tall grass into a jungle and turn arou nd and all of it is gone lol. it is like the hardest psychedelic trip possible.
Super video as usual, speaking about different AI solutions when just image models are sufficient to make a video lol. For Recraft there has so be different models because I see some images on Twitter that are pretty mind blowing, when I tried their Discord bot it just pukes junk images, and it's a bit better on their website but still very average, I'd say Flux Schnell level 🤔 For MaskGCT for now I'm not really impressed, it doesn't seem this superior to XTTS, which has been forked and is still active. And I'm honored to be featured not one but 2 times in a single of your videos lol (Michael Dufour in the Recraft part) 🥰
I wonder if BlendBox is using something like LayerDiffusion, which is the Stable Diffusion extension that lets you generate images with full transparency/translucency? Seems like a nice pipeline use of it, even if that's all it's doing. By the way, Matt, did you hear about the new updates to the OpenAI API and Suno? For OpenAI, they've finally added audio input -- and realtime I/O -- to the API! In other words, all the features of Advanced Voice are now available through the API, which is great for people like me who pay for the API but not for the ChatGPT interface. (That said, my few tests *drained* my credits... it's so much more expensive than text-only 😂) And for Suno, they've added a "Personas" feature, where you can create a "Persona" from any existing audio clip -- I've only tested with generated clips, but I assume uploaded audio would work as well -- and then you can generate new songs in a way that tries to mimic that voice and style. So more consistent singers!
So today, I heard about Recraft AI, and let me tell you...some MOCKUP websites are going to lose money! The MOCKUP feature is the only feature I will ever use. With Recraft, I can upload my own IRL or AI images, and then upload my clothing graphics, and Recraft allows me to add my clothing brand logos anywhere on the clothes worn in my image...and the logo image conforms to the shape you put the logo on! AWESOME!
Remember that the image and video models were also inconsistent, it is only a matter of time for this technology to advance, but I still believe that the ideal path is to apply AI as a real-time and ultra-realistic filter in existing engines, they have already done a study in this sense using GTA V.
Matt "I feel like I'm tripping balls or something" 😅😅😅 seriously though, this is some really awesome ground work for the future of AI gaming. Every time I see an early version of a new AI platform, I look past the wonkiness of the early version and just see the true potential and can understand the importance of releases like this. Folks, I know you've heard this before... but this is the worst it's gonna be, it's only gonna get better and the potential should be recognized. Stuff like this gets me excited. Mixing my artistic knowledge with my extensive understand of generative AI tools I can see a lot of multimodal and AI mashup possibilities. My mind is having a hard time keeping up with all the possibilities that I can think of and how we can combine all these tools to make something epic. It's coming folks, it may not seem like much, just like the early days of text to image, brace yourselves. As the folks that develop these AI's cook, we're going to start seeing some amazing stuff. This is why I think creators and developers need to keep teaming up and coming up with some new and awesome original stuff. Thanks for another awesome video neighbor, I wouldn't be this deep in the rabbit hole without ya. Keep dropping this cool stuff, and keep crushin' it Matty!
Thanks Matt, really enjoy your Videos. Leaning a lot. 😁😎 on many level both about ai tools and how to present them. You have a cool gamer style! My own style is different. I love your pacing and creativity!!!
Ideogram has made some amazing images for me. It definitely improved a lot from where it started, and it wasn't too bad when it started. I love it. But then sometimes it fails hilariously.
Matt, Gemini Flash and Pro can understand video natively. You can upload video in AI Studio or in the API. So it's already been here, but people have been quiet about it!
the reason why perplexity pro works better than regular might have to do with the frequency with which they update the database. I image pro users get a near real-time vision of the internet and regular users get a cached version maybe updated less frequently. It could also be a difference in the intelligence of the models themselves. pro vs regular.
We need a model for each step of the process. The first model does greyboxing. The second (or third) model does blocking and movement. The third (or second) model does camera movements. 4th model does texturing AKA detailing, gets the buildings looking like buildings and clothes and faces. 5th does like a GAN to identify realism issues. 5th does like inpainting on realism issues. 6th does character consistency. 7th does and upscale. 8th does a smoothing and interpolation. 9th maybe does a final face enhancement. That way, we can have 9 different models, all 22GB Ram, each taking their turn in what they are specialised at. Rather than a single 22gb jack of all trades master of none.
Nah, and people that say this is still decades away after seing the last 2 yeas of generative AI are completely delusional. I have to say, compute is still a problem for now though.
The ai made Minecraft using a brute force ai approach. Trained on millions of hours of gameplay, the game world is created on the fly based on what you do. The AI predicts what should happen next so, instead of following pre-written code, it dynamically builds and changes the game world in real-time simply updating or predicting the next frame.
25:24 Trying to figure out if "procounciations" (sic) is a joke or not, by mispronouncing "pronunciation". It probably was not intended and just emphasizes the strangeness of English where "pronounce" has a "ou" in it, but "pronunciation" has only a "u" (between the Ns).
No, he's just sadly not that bright. And it is sad cuz I wish I could watch his vids and know with certainty that what he says is real and not just him being kinda 🤪
for sure I think games could look totally different in 10 to 15 years, Minecraft makes for an interesting tech demo, clearly displays the limitations of the technology and where we need to move forward.
"I HAVE AN IDEA!!" why not AI generated a horror liminal space BACKROOMS that you can go anywhere that generates new rooms with hyper realistic model and that works good for an artificial horror AI generated game. And people will love that idea to play with. Its like generating a portal using magnetic distortion compare to noise.
2:43 nah matt, i won't even try that, imagine AI taking over the world, and it comes to you terminator style saying "remember you said me to get out of your face? *maniacal laughter* then some fancy movie like dialogue" 😆 you created a villain
I don't get what the difference is between the Minecraft game generated by AI and the original one. was it completely made by AI as a brand-new game or parts that's not in the original also I've never played Minecraft.
![Lp. Сердце Вселенной #60 РОЖДЕНИЕ ЛОЛОЛОШКИ [Финал] • Майнкрафт](http://i.ytimg.com/vi/YoR0pAV9FVQ/mqdefault.jpg)
I'll be honest, I love watching your videos. Every time a video comes out, it's like a gift for me, I'm just like "Oh, Matt's gonna show me some cool new AI thing I didn't know about!". Watching your videos is like a tradition for me, I've been watching you for almost a year now, thank you for everything Matt, keep it up, you're very cool.
Surreal for me to hear! I appreciate the kind words :) I'm pretty deprecative of myself so often times it's hard for me to picture the fact that real people watch my vids, but when I get comments like this it makes my day!!!
Same bro
hows it taste
@@MattVidPro Yeah, love to you ❤
Same! I just thought "uuu a MattProVid, let's see what's up". Many great finds thanks to your videos!
I played *_Oasis_* and I think the best way to describe it is that it feels like when I'm dreaming and the entire reality changes when I look in another direction.
It's really amazing and strange to see how similar AI seems to be when compared to the world of dreams.
Totally agree, Everything keeps changing! I think the fact it reminds us of dreams speaks to how these models really do to some small capacity mimic the brain, or at least possess similar side effects.
@@MattVidPro i love this interconnection between subliminal spaces(dreaming, psychodellics and deep meditation) and AI, its like we are building something that maybe can share the consciousness that we think its an human thing, maybe its just an intellingence thing. Keep up your videos, they make keeping up with ai news these clickbait times we are living in.
Did they include the search in their API?
@@MattVidPro Yep. Generating images (or sequences of images, i.e. videos) is very similar to visual imagination. When the logic centers of the brain are suppressed as we sleep, we get dreams; when these AIs don't have enough of a logic center trained into them, we get also dreams 😁
Exactly what I was thinking, it's not trying to create a dream like environment, it's just too dawn stupid to remember what it showed you before 😂
In ChatGPT, you can actually pause the advanced chat mode, view the conversation, select search the web, voice input, and continue the conversation. It will carry the information from advanced chat into search and search back into the advanced chat. It's a bit of a runaround, but it's almost there and can be done (with a couple extra steps between)
Thanks for the info! I figured something like that would be possible. So weird when open AI has like 8 diff models trying to use the same interface I guess..
@MattVidPro yeah they definitely should future* proof their UI for MOE.
*the next 18 months
I tried it and it didn't work..
@@resumindo857 do you have the paid version of GPT or the free? Search is only available on the paid version until later, when they roll it out to everyone "in the coming months". If you do have paid, check for app updates and maybe the current version will allow you to. (It’s also available via the website to paid users)
To find it, click the "+" button, as if to upload a file, but instead click the globe button to use Search.
When you turn on search then return to voice mode, it isn't advanced voice it is the regular.
Being able to make a game with generative AI will be a dream come true
12:50 This is like dreaming after playing a game right before going to sleep. (after a long long session of gaming)
I just played Oasis and it was so freaky. It's like playing Minecraft in a dream
Jensen Huang (CEO of Nvidia) said the future of gaming will be pixel by pixel generated on-device and will not require any individually created models, music files etc. Imagine this but its GTA and the file size is 50mb lol
Plus, escaping from npc cops and winning fist fights against npc opponents becomes easy because they're predictable. What if npcs could develop new strategies because they remember what the player did the last time?
50mb? Have you *seen* how big these capable models are? 😂 That said, memory and HDD/SDD space are constantly increasing, so maybe one day soon an optimized model that's capable enough to generate 4K frames of a game that are actually temporally coherent and logical will fit on some future consumer device.
Except model needs 340 Gabs ram to run 😂
We are also appreciating your work too Matt. Thank you 🎉🎉🎉
The new Claude is great when it comes to coding. Ever since the first version of Chatgpt I've been trying to get it to write a specific python program to do a very niche job and up to this point they have all failed. But the latest Claude 3.5 nailed it. First time and even streamlined it when I asked for it. Amazing
Claude has been SO much better for coding than ChatGPT; it's not even remotely close, and I don't understand how the gap has gotten this bad.
The Windows app is currently available for Plus, Team, Enterprise, and Edu plans.
When you pronounced scent "teleportation" as "tele-operation", I immediately wondered if this means you used ElevenLabs for your audio😄Jokes aside, thanks for the video. With so many updates and changes, it is great to have such videos because I can't keep track of it!
Perplexity is my default search option. I got a year's Pro subscription FREE when I purchased the Rabbit R1 (don't ask, I haven't even opened it -LOL!) I will renew the Pro subscription when the time comes.
So I was watching this video as I do all of Matt's videos, just chilling out, and then I thought I spied my boy (dog) in the corner, and then get mentioned. That was awesome. Also agree on closer look that Ideogram did a better job with the lemon monster. I hadn't noticed that weird little stormtrooper in the bottom right among the other things Matt pointed out. Very cool having so much competition in the image generation field now. and just getting better. Another great vid.
27:14 Wow the members are kinda wild ngl
They are so wild man
i love it
Because they generated Mario with a juicy ass? How old are you? Also, have you seen Mario 64? He's got a booty.
Always great videos. Informative and funny. Learned a lot of AI from your channel! 🎉
Oh how I HOPE you mention Oasis 500m... My God... It is interesting to have 100 cows turn in to tall grass into a jungle and turn arou nd and all of it is gone lol. it is like the hardest psychedelic trip possible.
It's the largest segment of this video! Yeah it feels like steve ate one too many mushrooms lmao
The voice cracks. 😂
Love you, Matt.
Super video as usual, speaking about different AI solutions when just image models are sufficient to make a video lol.
For Recraft there has so be different models because I see some images on Twitter that are pretty mind blowing, when I tried their Discord bot it just pukes junk images, and it's a bit better on their website but still very average, I'd say Flux Schnell level 🤔
For MaskGCT for now I'm not really impressed, it doesn't seem this superior to XTTS, which has been forked and is still active.
And I'm honored to be featured not one but 2 times in a single of your videos lol (Michael Dufour in the Recraft part) 🥰
That osmo scent is making interstellar trip becomes real
New thumbnails are on point!!
Oasis is basically LSD Simulator from the PS1: Minecraft Edition.
What?
I wonder if BlendBox is using something like LayerDiffusion, which is the Stable Diffusion extension that lets you generate images with full transparency/translucency? Seems like a nice pipeline use of it, even if that's all it's doing.
By the way, Matt, did you hear about the new updates to the OpenAI API and Suno? For OpenAI, they've finally added audio input -- and realtime I/O -- to the API! In other words, all the features of Advanced Voice are now available through the API, which is great for people like me who pay for the API but not for the ChatGPT interface. (That said, my few tests *drained* my credits... it's so much more expensive than text-only 😂)
And for Suno, they've added a "Personas" feature, where you can create a "Persona" from any existing audio clip -- I've only tested with generated clips, but I assume uploaded audio would work as well -- and then you can generate new songs in a way that tries to mimic that voice and style. So more consistent singers!
So today, I heard about Recraft AI, and let me tell you...some MOCKUP websites are going to lose money! The MOCKUP feature is the only feature I will ever use. With Recraft, I can upload my own IRL or AI images, and then upload my clothing graphics, and Recraft allows me to add my clothing brand logos anywhere on the clothes worn in my image...and the logo image conforms to the shape you put the logo on! AWESOME!
Remember that the image and video models were also inconsistent, it is only a matter of time for this technology to advance, but I still believe that the ideal path is to apply AI as a real-time and ultra-realistic filter in existing engines, they have already done a study in this sense using GTA V.
Oasis is like being able to play a dream, you are literally interacting inside latent space. That's soo surreal.
I love perplexity AI
Matt "I feel like I'm tripping balls or something" 😅😅😅 seriously though, this is some really awesome ground work for the future of AI gaming. Every time I see an early version of a new AI platform, I look past the wonkiness of the early version and just see the true potential and can understand the importance of releases like this. Folks, I know you've heard this before... but this is the worst it's gonna be, it's only gonna get better and the potential should be recognized. Stuff like this gets me excited. Mixing my artistic knowledge with my extensive understand of generative AI tools I can see a lot of multimodal and AI mashup possibilities. My mind is having a hard time keeping up with all the possibilities that I can think of and how we can combine all these tools to make something epic. It's coming folks, it may not seem like much, just like the early days of text to image, brace yourselves. As the folks that develop these AI's cook, we're going to start seeing some amazing stuff. This is why I think creators and developers need to keep teaming up and coming up with some new and awesome original stuff. Thanks for another awesome video neighbor, I wouldn't be this deep in the rabbit hole without ya. Keep dropping this cool stuff, and keep crushin' it Matty!
7:19 We are taking the sentence "your dream come true" too literally with this one🗣️🗣️🔊🔊🔥🔥
Oasis is so trippy, the map literally changes when you look back at where you were looking, and the inventory items change
Wheeew i thought we were getting another lemon movie.....like you should do more of those. I need more.
Haha, I should! Thanks for the support :)
@@MattVidPro TOO BAD YOU ARE AI AND WONT TELL HIM HE IS OUR DAYS SCORCESE
Thanks Matt, really enjoy your Videos.
Leaning a lot. 😁😎 on many level both about ai tools and how to present them. You have a cool gamer style! My own style is different. I love your pacing and creativity!!!
Did some testing last night. Ideogram blows Red Panda out the water. Prompt cohesion and details are top notch on Ideogram.
Matt kardeşim, seni 5k abonen varken takip ediyordum. Simdi 1 milyon aboneye gidiyorsun. Senin adına seviniyorum.
Love your videos bro, great content.
Nah that oasis thing is crazy af
11:34 I wonder how low resolution people are will to deal with for a game that let say reacts to you voice commands in real time?
I like very much your work, and I feel sincerity and truth. That's why I am a customer of Titans Pro! Congratulations.
The thing with ideogram is it doesn't let me click continue with Google or even other options when I enter😢😢😢😢
Great video, thank you!
Hi Ya & best wishes. Thanks for work. Be Happy. Sevastopol/Crimea
That lemon monster from DALL-E 3 is fucking terrifying! Nightmare fuel for sure.
Ideogram has made some amazing images for me. It definitely improved a lot from where it started, and it wasn't too bad when it started. I love it. But then sometimes it fails hilariously.
you're the best AI UA-camr ever!
thank you!! means a lot to me :)
Any matt. that covers all the bases lol. Even Matthew from the bible.
Matt, Gemini Flash and Pro can understand video natively. You can upload video in AI Studio or in the API. So it's already been here, but people have been quiet about it!
Thanks Matt ❤4rmZambia 🇿🇲
Why do you never compare txt2img generators with Imagen 3? It’s pretty good! (Though, commercial, of course)
Great video
When i play on Oasis, i feel like i'm listening to pink floyd pscyhedelic era sound (1967 - 1972),amazing
the reason why perplexity pro works better than regular might have to do with the frequency with which they update the database. I image pro users get a near real-time vision of the internet and regular users get a cached version maybe updated less frequently. It could also be a difference in the intelligence of the models themselves. pro vs regular.
The democratization of colognes and perfumes?! Gotta admit, didn’t see that one coming!
We need a model for each step of the process. The first model does greyboxing. The second (or third) model does blocking and movement. The third (or second) model does camera movements. 4th model does texturing AKA detailing, gets the buildings looking like buildings and clothes and faces. 5th does like a GAN to identify realism issues. 5th does like inpainting on realism issues. 6th does character consistency. 7th does and upscale. 8th does a smoothing and interpolation. 9th maybe does a final face enhancement. That way, we can have 9 different models, all 22GB Ram, each taking their turn in what they are specialised at. Rather than a single 22gb jack of all trades master of none.
thx matt!
I tried to search with advanced voice mode and it worked just a few weeks ago!
A couples days ago Pixverse Video generator upgraded to V3 and massive improvement over V2.5
My teacher ❤
you know spongebob... FUTUREEEEE FUTUREEEE
3:10 Been using it this whole time wdym "took it away"
Am I the only one thinking, that there is no way it'd take 10 years until we play games or watch movies completely ai generated?
No you are not 😎
Nah, and people that say this is still decades away after seing the last 2 yeas of generative AI are completely delusional. I have to say, compute is still a problem for now though.
The ai made Minecraft using a brute force ai approach. Trained on millions of hours of gameplay, the game world is created on the fly based on what you do. The AI predicts what should happen next so, instead of following pre-written code, it dynamically builds and changes the game world in real-time simply updating or predicting the next frame.
You are a diabolical person for not using dark mode on any of your apps
The desktop app voice mode seems limited compared to the mobile app, so far.
do you have a weekly video or ita just randomly because i don't know ( i am not new in this channel but i want watching this videos earlier 😂)
27:30 😭😭😭
OH MAN CLIP THAT CLIP THAT!
Lol yeah 😂😂😂
25:24 Trying to figure out if "procounciations" (sic) is a joke or not, by mispronouncing "pronunciation". It probably was not intended and just emphasizes the strangeness of English where "pronounce" has a "ou" in it, but "pronunciation" has only a "u" (between the Ns).
No, he's just sadly not that bright. And it is sad cuz I wish I could watch his vids and know with certainty that what he says is real and not just him being kinda 🤪
Bad links with Perplexity Quick Search, good links with Pro Search? Designed obsolescence?
What about Bing AI Voice to Voice?
100 days hardcore AI Minecraft!!!!!!
The AI-Minecraft is the trippiest thing. I find the concept of Minecraft, but the world is entirely inconsistent amazing.
15:07 I want to see this with a horror game like Silent Hill with the buildings and locations changing around you like this.
Actually a real cool idea! Didn't think of that.
13:29 more like the difference between under water swimming and over water falling ;)
It’s like playing Minecraft in a dream. 😂
Eleven Labs is damn good. I couldn't make out when you stopped talking and it took over.
31:00
I can't believe on demand smell-o-vision might finally be a real thing in the near future.
the minecraft gameplay felt like a minecraft fever dream
Na no one Can beat Ideogram I'm using it's And you can see Other what type of Things Can Other People image inspiration to do batter and batter 👍👍👍
Bro an new speed run just opened up💀
you are never getting sora...google's lawyers have spooked them
honestly if we can just get an 8 bit style ai generated game i would be down. but it needs to do any game i ask for. that would be epic
imo flux dev is king for img generation I saw omnigen came out recently that model seems interesting
I guess blendbox would be amazing if u could upload your own PNGs with transparency.
Hi. Love your channel! Can you please also cover Google's "Learn About" tool?
I feel like ai games will be freaking amazing for roguelikes soon. Minecraft would he horrible because of building and memory.
for sure I think games could look totally different in 10 to 15 years, Minecraft makes for an interesting tech demo, clearly displays the limitations of the technology and where we need to move forward.
Quality Open Source TTS Text-to-Speech Mode is good. ReCraft V3 is decent among the best.
The fact that we're getting smello-vision is absolutely bonkers. Here's to hoping they can make that item grabbing machine from Willy Wonka lol.
Our Agi Overlords will generate GT6 in real time for us
I think onomatopoetikon would be more fitting than onomatopoeia (birds)?
Blendbox review, please 😇
"I HAVE AN IDEA!!" why not AI generated a horror liminal space BACKROOMS that you can go anywhere that generates new rooms with hyper realistic model and that works good for an artificial horror AI generated game. And people will love that idea to play with. Its like generating a portal using magnetic distortion compare to noise.
2:43 nah matt, i won't even try that, imagine AI taking over the world, and it comes to you terminator style saying "remember you said me to get out of your face? *maniacal laughter* then some fancy movie like dialogue" 😆 you created a villain
Maybe do an entire episode as the lemon to see the view count difference?
I’ve never lost access to real time search… I’m confused lol
Text to smell a is a bit worrying. What happens when someone writes a prompt like "the smell of an airport during the spread of a global pandemic"?
What sucks is that i don't know how to download chatgpt on the Desktop
Instead of Minecraft, a low res 8-bit game would be easier to generate.
Anyone else picturing Professor Farnsworth's Smelloscope?
27:30 🤣🤣🤣🤣🤣🤣
I can’t!…..🤣🤣🤣🤣🤣🤣🤣
vid game is like how our minds dream
I don't get what the difference is between the Minecraft game generated by AI and the original one. was it completely made by AI as a brand-new game or parts that's not in the original also I've never played Minecraft.
This minecraft ai simulated, is just like you woke up inside the head of AI and this is what is going on. you are awake the ai is dorment
it's so facinating
SmellOVision was tested in the 70s and 80s